The Career of Mobbing: Emergence, Transformation, and Utilisation of a New Concept

Home , Dan Olweus, Exaggeration, Heinz Leymann

Rapport nr 29, 2008

OLA AGEVALL

Institutionen för samhällsvetenskap ______

VÄXJÖ UNIVERSITET

RAPPORTSERIEN Ges ut av Institutionen för samhällsvetenskap vid Växjö universitet. Serien innehåller rapporter och uppsatser författade av lärare och studenter vid institutionen.

BESTÄLLNINGAR OCH FÖRFRÅGNINGAR Institutionen för samhällsvetenskap Forskningssekreterare Växjö universitet 351 95 VÄXJÖ

2 Abstract: Concepts play a part in any scientific endeavour. They simultaneously serve as receptacles for previously attained knowledge, as vehicles of scientific and di- dactic communication, and as building blocks in the construction of new descrip- tive and explanatory schemes. In this sense, concepts are central to the cognitive organisation of scientific disciplines. But disciplines also differ from one another, and not just in terms of the stock of concepts deemed necessary for internal scientific communication but also, for example, in terms of the logic of conceptual change, the relative influx of connotations, and the degree of universal adherence to specific definitions. One crucial characteristic of the social sciences is that they share part of their vocabulary with the lay people they study. In some cases, social scientific terminology has trickled into everyday speech; in other cases the social sciences have proceeded by refining and redefining locutions found in natural languages. The aim of this book is to study empirically, by means of a detailed case study, the mechanisms that shape concept formation in the social sciences. It focuses on the history of the concept of ‘mobbing’, with special reference to the interaction between science and lay conceptions. This condensed conceptual history shows that the public discourse, that was instrumental in the diffusion and entrenchment of the notion of mobbing in the early 1970s, significantly shaped the career, contents, and uses of the corresponding scientific concept.

Keywords: Mobbing, Bullying, Concept formation and transformation, Issue- symbiosis, Conceptual contagion, Conceptual history, Sociology of science.

3 Table of Contents

1. Introduction...... 5 2. Enigmas in the conceptual history of ‘mobbing’ ...... 7 3. Between journalism and ethology...... 10 4. Issue-symbiosis, conceptual contagion, and institutional backing...... 18 5. The early image of mobbing: an analytical model ...... 25 6. Mobbing and the scientific literature ...... 27 6.1. Dan Olweus and mobbing cum aggression...... 29 6.2. The meaning of ‘mobbing’ ...... 34 6.3. Olweus’ empirical results and theoretical framework ...... 40 7. Conceptual entrepreneurialism and embedded discursive remnants: an analytical summary ...... 47 8. Coda: a note on ’mobbing’ since the mid 1980s...... 52 References...... 63

4 1. Introduction

It is a commonplace that the social sciences share the vocabulary of those lay people who are their objects of inquiry. Save for a handsome amount of technical or quasi-technical terms, we deal in the same ordinary language categories as the people we interview or the bureaucracies we study. From time to time, the relation between ordinary language and social science terminology has been turned into an object of inquiry. In the early 1960s, Serge Moscovici (1961) studied patterns of distortion that recur when social scientific concepts trickle into ordinary language and are used by laymen. In the 1970s, the conservative German phi- losophers of the Tendenzwende argued that social scientific terminology was pol- luting everyday speech. But terminology does not flow in one direction only. Words with an extra-scientific pedigree are often picked up, modified, and turned into scientific concepts by the social sciences. The aim of this essay is to begin to charter, in a detailed case study, how such processes work.1 The concept of ‘mobbing’, whose evolution we will follow in this essay, has some properties that make it into a suitable laboratory for that purpose. First of all, it has an identifiable and quite recent date of birth in ordinary language. Se- cond, a lot has happened to it in the course of just a few years, from November 1969 to 1973. Although I will make considerable excursions outside of this time span, the sequence of events that is played out in this condensed period is central to an understanding of subsequent events: the timing, manner, and context of the birth of ‘mobbing’ contributed to the shaping and entrenchment of a dominant strand of research in the field. Third, while ‘mobbing’ as a concept began as a local Swedish affair, it has since become staple goods in international social scientific discourse. What I offer is less than a full account of how the term traver- sed national borders, but there is enough material here to give an idea of how it

––––––––– 1 A sibling to this essay – a history of the concept of stereotypes – is under preparation.

5 happened. A relatively simple and condensed story, ‘mobbing’ is ideally suited for extracting some basic ideas about conceptual careers in the social sciences. In the pages to come, we will follow its meanderings from lay conceptions to scientific concept and back again. It is a tangled story, and an intriguing topic for the sociology of knowledge and science, involving complex interactions between science, media, school, new and old associations, law, and public administration.

6 2. Enigmas in the conceptual history of ‘mobbing’

In 1969, a general physician from Sweden published an article on ‘apartheid’ in a small left-liberal journal. Contrary to what you would expect from the title, the article is not primarily about South African segregationist politics. The author is the concerned father of a black adoptee boy, and he is writing down his reflections on incidents of exclusion and harassment that his son had experienced. Moreover, he does not describe his son’s predicament in terms of racism. The everyday exclusion and harassment he is subjected to, is portrayed as a sort of group violence directed against a single, singled-out individual. The fact that his son is black does matter, for it sets him apart from his social environment; it makes him conspicuous, and so increases the likelihood that he will be targeted for exclusionary and harassing treatment. But other ‘deviant traits’ could have the same effect, and in this sense it is arbitrary that his son happens to be black. Peter-Paul Heinemann, the author of the article, was concerned not with racism but with a phenomenon referred to as ‘mobbing’. The word itself was a ne- ologism, derived from the English verb ‘to mob’, and inspired by Konrad Lo- renz’ work on mob behaviour in animals. No Swedish expression, Heinemann thought, captured what he had in mind: a group of people acting in concert to ex- clude and/or harass a single individual. His terminological deliberations have their ironies, for in English texts the concept ‘mobbing’ is usually considered cumbersome and is translated back into English as ‘harassment’ or ‘bullying’. What comes out as startling, however, is the fact that it has been translated back into English. Some thirty-five years after the publication of Heinemann’s article, in a small Swedish language journal, the concept has trickled into Ang- lophone scholarship. In Sweden it has acquired legal significance, thus spawning

7 a flood of documents to regulate liability and due process. The word ‘mobbing’ itself has been taken up in foreign languages, e.g. German and Norwegian, and is arguably one of Sweden’s main linguistic exports in modern times. A pivotal event in this history was the publication of psychologist Dan Olwe- us’ book on mobbing in schools in 1973, some three and a half years after the appearance of Peter-Paul Heinemann’s article on apartheid (Olweus 1973). It was the first in what would prove to be a long series of publications on the topic, which in due time made Olweus one of the most cited Swedish scholars in the field of social science.2 Needless to say, Olweus and others have since refined the approach. But the questions posed in that first book, the definitions it suggests, the methodology it employs, and the theoretical outlook it draws upon set the agenda for future research. Olweus’ first book in the field was no mean achievement. It reported results from five empirical investigations, comprising data on nearly one thousand twelve to sixteen year old boys in Solna and Stockholm, and using a variety of quantitative methods to assess prevalence and test hypotheses. It was also the first scholarly book on mobbing. Yet in the preface the author explains that the intended audience for the book is the general reader. The reason why matters of method and theory nevertheless loom large in the book, he explains, is that the conclusions he had arrived at either ‘questioned or explicitly contradicted current conceptions of the problem’ (Olweus, 1973:5). He was looking to set the record straight, to uproot unfounded but thoroughly entrenched conceptions, and at the same time launch his own explanatory model. In retrospect, there is something enigmatic about Olweus’ declaration of in- tent. Considering that only three and a half years had lapsed since the appearance of Peter-Paul Heinemann’s article on apartheid, how was it possible for erroneo- us ideas about mobbing to have become so thoroughly entrenched? Whence did these deeply rooted ideas come from, and what were they? And how was it pos- ––––––––– 2 As of June 21 2006, the Social Science Citation index lists 2036 articles quoting Olweus’ works. As a point of comparison, we may note that Göran Therborn, by all standards an internationally re- nowned Swedish scholar, was quoted in 999 articles in the same period, i.e. less than half of the quotations harvested by Olweus. Indeed, a sociological classic such as Émile Durkheim just barely

8 sible for Olweus to initiate, conduct, write up, and publish his research within that same and very short time-span? As we shall see, the answers to these questions are linked both internally and with subsequent developments in an emer- ging field of research.

surpasses Olweus’ record, scoring 2255 quoting articles. In 1994, The Times dubbed Olweus the world-leading authority on bullying (Forsman, 2003:98).

9 3. Between journalism and ethology

Had Heinemann’s infant concept remained a matter for the small readership of Liberal Debatt, it would certainly not have made much splash. This was not to be the case. Two journalists – Anna-Maria Hagerfors and Birgitta Nyblom – took notice of his article and made it their point of departure for a series of articles. The articles were published in Dagens Nyheter, a large Swedish daily newspaper with national coverage, during November and December 1969. By the time of the final article in the series, the word mobbing had already become established in everyday speech.3 The articles triggered an avalanche. Swedish Broadcasting Company broadcasted programs devoted to mobbing, as did the Swedish television.4 Other newspapers followed suit.5

––––––––– 3 Olav Panelius & Torsten Steinby (1970) lists mobbing as a vogue word already in 1970. Since the word remained in frequent use, the Swedish Language Council decided, in 1972, to reform the spelling from English ‘mobbing’ to Swedish ’mobbning’. 4 I have only found one indexed reference to mobbing in the archives in the months after the concept’s first appearance: a newspaper article, registered in SAOB Arkiv, notes that mobbing was the theme in the radio show Familjespegeln 15 March 1970. Radions kortkatalog, which contains older materials not catalogued at Statens ljud- och bildarkiv, is not complete and has no entry for this programme. It is likely, however, that mobbing was taken up in Familjespegeln on more than one occasion. In a retrospective interview, Karin Wilhelmsson – the journalist in charge – recalls that her show was instrumental in diffusing the concept (cf. interview in Vår grundade mening 10 Au- gust 2006). Although this may be an exaggeration, Familjespegeln did pick up mobbing later on, e.g. 24 November 1971 (Radiohusets kortkatalog). A few more early radio programmes have been preserved where mobbing was discussed, e.g. Du och de andra 20 August 1973 and Det här med barn 12 March 1973. An interesting research endeavour, which I have not undertaken, would be to investigate where mobbing is mentioned in radio without being the focal theme of the programme. My guess, based on the argument below, would be that it recurs in connection with discussions of school size, physical deviance, immigration, adoption, and urbanisation. Mobbing was also taken up in television. In 1969, Sweden got its second TV channel – TV2 – often referred to as the ‘new radical television channel’. 21 December that same year, the TV2 show Det händer nu was devoted to mobbing. As in the case of radio, Statens ljud- och bildarkiv does not systematically catalogue such early materials. But it does contain a couple of more entries between 1969 and 1973. 11 december 1970, Tv-nytt broadcasted from a debate – organised by the Swedish Red Cross and the Immigrant Women’s Association, and discussed in more detail below – about ‘the deviant children and the mobbing against them’. And 14 October 1973, the TV2

10 Journalists were not the only ones to show interest in the phenomenon. During this early phase of the diffusion process, we find letters-to-the-editors picking up on the subject.6 Soon there were also responses from and within organisations and institutions. Before Olweus’ research had been conducted and published, theatre plays had been written and toured the nation’s schools7, the Immigrant Women’s Association had become involved and had among other things arranged a public hearing8, the national PTA organisation had taken up the issue9,

programme Knuff devoted a programme to mobbing. Cf. Röster i radio och TV (1973), for a presen- tation of the programme. See also the following programs, all broadcasted between 1969 and 1973: Skol-TV Jag själv? 2.Hackkycklingen – vad är den bra för? 4 October 1970; Insändaren 1 Novem- ber 1970; Rapport 2 November 1972; Filma Själv: Verkligheten har många ansikten 3 May 1973; Våld och nöd: Skol-TV del 1: vad är våld 16 April 1973; Aktuellt 27 September 1973; Rapport 4 October 1973; Du och de andra 5 October 1973; 5 As evidence we may quote the following articles, most of them found in SAOB Arkiv: Svenska Dagbladet 23 November 1969; Svenska Dagbladet 15 March 1970; Svenska Dagbladet 10 March 1971; Svenska Dagbladet 18 december 1971; Svenska Dagbladet 12 February 1972; Svenska Dag- bladet 21 September 1972; Svenska Dagbladet 17 April 1972; Svenska Dagbladet 16 June 1973; Svenska Dagbladet 15 October 1973; Göteborgs Handels- och Sjöfarts-Tidning 13 July 1971; Göteborgs Handels- och Sjöfarts-Tidning 28 October 1972; Östgöta Correspondenten 12 February 1970; Östgöta Correspondenten 6 May 1970; Östgöta Correspondenten 16 December 1970. Dagens Nyheter continued to publish articles on mobbing. Moreover, the list of articles above is far from complete. Large Swedish newspapers such as Aftonbladet and Expressen are not included. A glance through Svenskt Tidskriftsindex for this period reveals that mobbing became an index word in 1970. In the period between 1969 and 1973, it contains the following articles: Pall, Arne (1970) ”Mobbing”, Psykisk hälsa (11), h.3, s.181-190; Toijer-Nilsson, Ying (1970) ”Mobbing kallas det”, Barn i hem skola och samhälle, 24, h.9, s.33-35; Staaf-Busch, Dagny (1971) ”Porträtt av en mobbad”, Vi föräldrar, 26, h.11, s.36-38; Vilson, Gunnar (1971) ”Vad mobbing är för något”, Barn i hem skola och samhälle, 25, h.6, s.24-27; ”Mitt barn blev mobbat” (1971), Barn i hem skola och samhälle, h.2, s.9-13; Langton, Kaisa (1972) ”Bäddat för mobbning”, Barn i hem skola och samhälle, 26, h.8, s.31-32, 44; Renström, Christina (1972) ”Vad är mobbning och vem blir mobbad?”, Kommunal skoltidning: tidskrift för skolstyrelser och skolledare, 41, h.9, s.386-390; Vogel, Viveka (1972) ”Vi måste alla ta ansvar för mobbning!”, Vi föräldrar, 5, h.10, s.44-45, 81; Johan- nesson, Inger (1973) ”Dan Olweus: Hackkycklingar och översittare: Mobbning i skolan”, Psykisk hälsa, 14, h.4, s.360-370; Nilsson, Bertil G. (1973) ”Varför blir vissa barn mobbade?”, Försäk- ringstidningen, 28, h.5, s.4-7; Vilson, Gunnar (1973) ”Trygghet och säkerhet eliminerar mobbningen”, Svenskbygden: organ för finlandssvenskt bildningsarbete, 52, h.6, s.88.89. Once again, it must be emphasised that the enumeration is far from exhaustive. 6 See, for example, Svenska Dagbladet 23 November 1969. 7 Ingrid Sjöstrand, famous in Sweden as the author of the children’s book Kalle Vrånglebäck (1968), wrote a play on mobbing, Är du rädd för Uffe? Han är också rädd. We discuss the contents of this play further on. At this point, we emphasise the very brief time-span between the first appearance of mobbing and the first play, and it suffices to note here that the theatre group Grupp Q first staged it in February 1970 (Dagens Nyheter 5 February 1970; Svenska Dagbladet 15 March 1970). A couple of months later, newspapers reported that school children had authored their own play about mobbing (Östgöta Correspondenten 6 May 1970). 8 See section three below for a discussion of one of these events. Another example is a panel discussion that was organised by the Immigrant Women’s Association 21 February 1970 (cf, Dagens Ny- heter 22 February 1970).

11 questions and motions had been raised in the Swedish Riksdag10, and the Swe- dish Board of Education had constructed a teacher’s manual and recommended it to be used in teachers’ seminars in all local schools.11 With so many vehicles of diffusion, the next question to ask is what was being diffused. In one sense, it was obviously Peter-Paul Heinemann’s conception of mobbing, as expressed in the article on apartheid, which was transported through these channels. Yet Heinemann did not provide a definition of the phenomenon he named. The two principal anchor points of the concept were instead the examples he offered and his general reference to Konrad Lorenz’s work in ethology. Both decisively influenced the development of the concept. In relying on example rather than definition, the class denoted by the term ‘mobbing’ was open-ended and fuzzy. Heinemann’s examples had struck a chord

9 Svenska Dagbladet 17 April 1972. 10 See entries in Register till Riksdagstrycket 1971-1980. Moreover, the directives to the state commission on the internal work of the schools, issued in 1970, contains the following specification: One group of pupils, whose difficulties in school is not due to a lack of motivation, are those who experience so-called mobbing because they are perceived to be deviant from their peers. Attention must be paid to this type of persecution, both regarding those pupils who are the victims and those who participate in the mobbing activities. It is also necessary to create a space for discussing these issues in class, and to instil in the pupils an understanding for different lifestyles and behaviours. (SOU 1974:53, p.65.) Despite the directive, however, the commission did not specifically address the mobbing question. Nevertheless, politicians were deeply involved in discussions about mobbing. Confer also the request – from a Social Democrat member of Stockholm’s Municipal Council – that the Swedish Board of Education should make sure that schoolteachers get a proper education on mobbing (Dagens Nyheter 18 November 1969). 11 Cf. Skolöverstyrelsen (1973) Mobbning i skolan: Handledarhäfte, and Skolöverstyrelsen (1973) Mobbning i skolan: Elevhäfte. The materials compiled by the Swedish Board of Education were not only mandatory for all school personnel, it also gave quite detailed instructions for what this in- house training was to be conducted. An excerpt from the table of contents gives a flavour of the level of detail involved: Page Headline Time slot 6 We are different 30 minutes 9 Why do you get mobbed? 20 minutes 10 The mobbed teacher 90 minutes 11 Our fundamental needs 30 minutes … … … 14 Why do we mob? 20 minutes 16 Five cases and some assignments 95 minutes 20 An acute mobbing situation 30 minutes The authority of the Swedish Board of Education to have every teacher participating in the same training no doubt contributed to the diffusion of the concept. But there were other centralised channels that provided for a swift diffusion among schoolteachers. The Swedish library service has long supplied the schools with a continuous selection of newspaper articles deemed to be relevant to the

12 in the public mind, evoking memories of past, and images of present, personal experiences. This is evident from the fact that journalists, organisational representatives, and the general public readily supplied additional examples. But rather than instantiating a preconceived and explicitly circumscribed category of actions, these additions proceeded by analogy with the corpus of existing examples. Each addition would then have to be similar to the available body of examples in some respect, but all need not pick up on the same aspects. And as examples accumulate, the corpus of examples from which to draw new analogies is liable to become increasingly diversified. Inch by inch, such processes work to al- ter the extension of the concept, thereby affecting its implicit intension. In this sense, the concept ‘mobbing’ had multiple authors, each contributor adding not only to the stock of examples but also to the potential meaning of the term. We shall return shortly to the contents of these various contributions. Heinemann’s reference to Konrad Lorenz’s work on aggressive behaviour in animals was equally important. Lorenz’s popular book Das sogennante Böse: Zur Naturgeschichte der Aggression (1963) had appeared in a Swedish translation in 1967, with a second edition appearing in 1968. This was Heinemann’s immediate source of inspiration, and the narrative of his article on apartheid phenomena wedded everyday examples of harassment and exclusion to a vocabulary taken over from ethology. The latter had primarily reserved it for cases where a flock of herbivorous animals make a concerted, pre-emptive attack on a predator.12 But on one occasion Lorenz did use the term somewhat more figuratively and applied it to human behaviour:

Aggression elicited by any deviation from a group's characteristic man- ners and mannerisms forces all its members into a strictly uniform obser- vance of these norms of social behaviour. The nonconformist is discrimi- nated against as an 'outsider' and, in primitive groups, for which school classes or small military units serve as good examples, he is mobbed in the most cruel manner. Any university teacher who has children and has held positions in different parts of a country, has had occasion to observe the amazing speed with which a child acquires the local dialect spoken in

school world (the so-called Skolans artikelservice). Thus, when newspapers started to write about mobbing, teachers would get access to a condensed selection of articles. 12 Cf. Konrad Lorenz (2002) On Aggression, pp. 23-25.

13 the region where it has to go to school. It has to, in order not to be mobbed by its schoolfellows, while at home it retains the dialect of the family group. Characteristically, it is very difficult to prevail on such a child to speak, in the family circle, the 'foreign language' learned at school, for instance by reciting a poem. I believe that the clandestine membership of any other than the family group is felt to be treacherous by young children.13

The elements of this passage are echoed, almost verbatim, both in Heinemann’s own examples and in subsequent accounts in the newspapers. There is no question, then, that Heinemann’s concept of mobbing enveloped Konrad Lorenz’s point of view. This had two consequences for the conception and reception of mobbing. First, it contributed to naturalising the phenomenon. Mobbing was simply a subtype of aggression. Aggression, in turn, was governed by natural laws, the description of which was the province of ethology and psychology. ‘Mobbing’ thus acquired an air of scientific legitimacy, much facilitating its reception among the lay public, and received a scientific legacy that loaded the dice in favour of particular scientific disciplines.14 Second, the imagery evoked by Lorenz’s mobbing animals restrained and directed the analogies employed by those laymen, journalists, and organisational representatives who subsequently added to the corpus of examples of mobbing. This last point brings us back to the contents of these contributions. When Peter-Paul Heinemann’s reflections first ascended into the public sphe- re in 1969, he provided three things: a name, a set of examples, and an interpre- tive background in ethology. Naming provided a preliminary sense of unity. It created a new category comprising a set of behaviours that, until then, either had gone entirely unperceived or were not thought of as internally related.15 But in

––––––––– 13 Lorenz (2002) On Aggression, p.76. Peter-Paul Heinemann explicitly refers the reader to this passage in his 1972 book on mobbing. See Heinemann (1972) Mobbning: Gruppvåld bland barn och vuxna, p.7. 14 Konrad Lorenz received the Nobel Prize in medicine 1973, sharing it with Nikolaas Tinbergen and Karl von Frisch, but in 1969 he was already a scientific celebrity representing a hard science. 15 The following observations were made in a newspaper article, published in Göteborgs Handels- och Sjöfarts-Tidning 28 October 1972: It is only a couple of years ago that the English loanword mobbing – which today has received a reformed Swedish spelling – was added to our language. The phenomenon it refers to is age- old. But until recently, the Swedish language only had words for its different forms of expression; there was no word for the concept as a whole. It is striking how much attention the prob-

14 the absence of a definition – and, needless to say, in the absence of empirical knowledge as to the nature, extent, causes, and consequences of the phenomenon thus named – that set of behaviours remained undetermined. In the years to come, a variety of agents, some of which are mentioned above, were to contribute their own examples and understandings of ‘mobbing’, thereby gradually fle- shing out the concept. As the next few examples will show, the view of Konrad Lorenz, as mediated through Peter-Paul Heinemann, was a central, but by no means the only, source in this build-up. 13 November 1969, headlines in Dagens Nyheter announced: ‘Fatso, limp, Finnish bastard – it easily ends in racial hatred’, ‘Mobbing is a prelude to apartheid in Sweden’, ‘Peer groups pose the greatest danger: the number of children has a strong impact’, and – three days later – ‘The harsh world of the schoolyard: deviant children are the whipping boys of their classes’. Thus began the series of articles that established mobbing in the collective consciousness (and conscience) in Sweden. What is striking about these headlines is that they are exclama- tion rather than question marks. They give the impression that we already knew quite a lot about mobbing. The ultimate consequence of mobbing is racial hatred and apartheid; the victims of mobbing are physically deviant from their peers, immigrant or adoptee children being especially likely to be affected; large school classes facilitate mobbing. A closer inspection of the articles gives a good idea of the curious blend of humanist concern over racial inequality and ethological argument that went into this first conception of mobbing:

Peter Paul Heinemann … has investigated the phenomenon in Liberal De- batt’s issue on apartheid, and suggested that the word mobbing should be used in Sweden to describe human behaviour. The word is taken over from ethology, especially from Konrad Lorenz, who has studied mobbing in animals, primarily birds: A large group of birds attack and drive away or even kill a bird of another species. Sometimes the target is of the same

lem has been accorded after this linguistic deficiency was repaired. Semantic innovations certainly don’t lack practical importance. The view that mobbing was a new name for an old phenomenon is found elsewhere. See, for example, Svenska Dagbladet 15 March 1970; Christina Renström (1972) ’Vad är mobbning och vem blir mobbad?’, Kommunal skoltidning: tidskrift för skolstyrelser och skolledare, I:9, p. 386; Barbro Goldinger (1974) ’Att övervinna mobbningen är inte så lätt som Olweus tror’, Lärartidningen: svensk skoltidning, 30-31, p. 11

15 species as the attackers. Applied to humans, we can compare this with lynchings and pogromes. […] I am the father of a seven-year old Negro boy, he[Heinemann] says. During his lifetime I have been convinced that the mechanisms of apartheid are alive and well in our country. I share this experience with all parents whose children strongly deviate from their peers.16

Heinemann took over, not only the terminology, but also the basic explanatory mechanisms, from Konrad Lorenz. In Lorenz’s view, aggression is an innate in- stinct or drive. He holds that there is a specific aggressive energy that is continually operative in the organism. According to that energy model of motivation, ‘this leads to the building up of tensions that the individual tries to discharge in one form or another’ (Olweus, 1970:5; Lorenz, 2002). If tension is built up without opportunity of release, the individual will eventually actively seek some form of tension reduction. This is exactly how Heinemann, in the first article in Dagens Nyheter, described the mechanism that triggers mobbing:

According to Lorenz’s book On aggression, violence and pleasurable enthusiasm derive from the same source of aggression and are found to be interchangeable. If the enthusiasm of children’s play is disturbed, it becomes violent. Dr. Heinemann’s conclusion is that child mobbing is a phenomenon of a violent nature, which occurs when such reorientations are not accompani- ed by an acceptable outlet to their aggression or enthusiasm. Children in suburbia and in the big cities lack the escape opportunities available to those who live in the countryside or in sparsely populated areas. This is a sure guarantee that there will be easily available victims of mobbing: They simply cannot escape. - Mobbing increases in proportion to the size and density of cities, according to his experience. The children of suburbia are under-stimulated from lack of freedom to move and act. At the same time, they are over- stimulated because of the large number of children in proportion to the space available for them to play upon. They are forced upon each other to an extent that they are unable to cope with. Their most common game is to wait around for something to happen. - When you pass through such a crowd of kids with a Negro or Korean child by your hand, they immediately start to mob. If the victim is alone, he is assaulted. A poor lightning conductor has happened to pass by, and

––––––––– 16 Dagens Nyheter, 13 November 1969.

16 has done its service. The more afraid he is, the more powerful is the conductor effect. - This has nothing to do with racism, he points out. Not yet, at least.17

We recognise the energy metaphor from Lorenz. Aggressive energy has been stored up in the crowd, as in a mounting thunderstorm, and the deviant boy pas- sing by unleashes their mobbing behaviour the same way a lightning conductor attracts a bolt of lightning. We also recognise the original ethological meaning of ‘mobbing’, for the phenomenon is portrayed as a case of group violence, of ‘all against one’, analogous to a flock of herbivore birds attacking a lone predator. This was potent imagery, and it was taken up in subsequent lay contributions in the following years.18 Mobbing was synonymous with ‘group violence’ – this term, too, became a vogue word.19 Heinemann himself emphasised that mobbing was tantamount to all-against-one situations (Heinemann, 1972). And when newspapers needed illustrations to their articles on mobbing, the motive of the arranged photo-shoots often neatly conformed to the ethological image: one boy is lying on the ground, a group of several boys standing in a circle around him.20

––––––––– 17 Dagens Nyheter, 13 November 1969. 18 The following instruction is taken from educational material, ‘Mobbning i skolan’, compiled by the Swedish Board of Education in 1973 and intended for use at in-service training days for all teachers in Sweden: The description of the park [in the student handout] where the children are playing conjures up the image of those poor playgrounds that we, all too often, offer our children in metropolitan areas. There is asphalt and thorny bushes, half-heartedly embellished with a sandpit in the centre. The lack of stimulating activities accumulates the natural need for excitement in each of the kids, and in unfortunate cases, this need can be the factor that triggers mobbing behaviour. Now discuss and assess the equipment available in your own schoolyard, and in neighbouring schoolyards or playgrounds you are familiar with! Give suggestions as to how they could be given a different and better design! Is centralised planning too strict, and does it inhibit individual spontaneity? (Skolöverstyrelsen, 1973a:16) The diffusion of this in-house training manual is discussed in note 11 above. 19 Cf. “Svenska Modeord“, Dagens Nyheter 5 February 1970. 20 See, for example, Dagens Nyheter 13 November, Ying Toijer-Nilsson’s article ”Mobbing kallas det”, Barn i hem skola och samhälle, or the picture on pages 55 and 174-75 in Olweus (1973). Cf. also the film sequence in the TV programme Våld och nöd: Skol-TV del 1: vad är våld 16 April 1973, where a girl is mobbed in a school corridor.

17 4. Issue-symbiosis, conceptual contagion, and institutional backing

Heinemann’s notion that deviant children are the victims of mobbing also appears to have a Lorenzian pedigree. It may not follow from the image of mobbing birds, but Konrad Lorenz had explicitly said that a group of internally similar children would be prone to punish deviators. But Heinemann’s application of the concept also contained elements that were of his own device. While aggressive instincts or drives were made out to be the final cause of mobbing, the immediate cause was conditions of life in suburbia and the big cities. According to Heine- mann’s experience – no data was available – mobbing was more common in the big cities than in the countryside. And further on in the interview, Heinemann extended the argument to school size: ‘He reminds us that the schools of tomor- row that we are building today will admit over one thousand pupils; in those schools, children can do outrage unto each other without anyone taking notice.’21 Even if Konrad Lorenz’s views formed the basis of Heinemann’s syllogism, Lo- renz himself was reluctant to go along with his conclusions. Konrad and Margre- the Lorenz happened to be in Sweden in November 1969, for the Nobel sympo- sium, and journalists from Dagens Nyheter took the opportunity to ask him about the effects of school size on mobbing:

- A group of internally similar children will always react upon a stranger, he [Lorenz] says. It does not necessarily have anything to do with density. Child mobbing occurs in small schools as well. If the strange child comes close to the internally similar group of children, they will be aided by larger numbers of children. In this regard, a densely populated environment will matter.22

––––––––– 21 Dagens Nyheter, 13 November 1969. 22 Dagens Nyheter, 13 November 1969.

18 Despite Lorenz’s hesitant and largely negative reply, the headline of the article summarised his views as ‘Peer groups pose the greatest danger: the number of children has a strong impact’. The journalists had sided with Heinemann’s extended argument. This aligned mobbing with other perceived problems of the day, such as the adverse effects of large schools and the alienation thought to be inherent in metropolitan life. Both were matters of concern and targets of criti- cism in Sweden by the late 1960s.23 Whereas Lorenz bestowed the concept of mobbing with legitimacy and provided it with a core meaning, elaborations of it did not keep within the bounds of ethological imagery. Heinemann’s proposition that school size and life in metropolis are the culprits, for example, locked mobbing into a nexus of contemporary debates and issues. Those who already abhorred urbanisation, or considered the new large schools to be an evil, were thus reinforced in their beliefs and provided with a set of new arguments. Conversely, the concept of mobbing gained addi-

––––––––– 23 The two issues are separate but historically intertwined. Urbanisation came late in Sweden. It escalated on a large scale in the 1960s. Employment opportunities were increasingly located to the larger cities, and the equitable wage policy of the Social Democratic party reinforced the push towards migration. Regional planning was introduced to soften the blow for municipalities in rural areas, but the net effect was nevertheless a decisive pull towards urbanisation. It was a major shift, perceived by many to put severe strain on the social fabric. By the end of the 1960s and in the 1970s, there was a rumbling discontent in some quarters. Scepticism about urbanisation took different forms. On the one hand, there was fear that metropolitan existence would be anonymous, deperson- alised and alienating. On the other hand, people in the depopulating areas sensed that migration threatened the existence of their villages, economy, and way of life. Holmberg (1998) discusses in detail the widespread popular critique of urbanisation and urban civilisation that emerged in response to these changes. The notion that mobbing was produced by life in suburbia and metropolis harmonised perfectly well with this critical undercurrent. These processes were linked to the issue of school size. In 1962, there was a reform to re- duce the number of municipalities - aiming, among other things, to create economically viable administrative units. But as Ulf P. Lundgren observes: Conditions were altered by developments in the 1960s. The population structure changed. The migration flow went from rural to urban areas, where there were employment opportunities. This altered the tax base, and the system of compensation between municipalities did not turn out as expected. Population growth in regional centres and larger cities required new school units, which in turn – as a consequence of the school structure – meant large school units. (Lundgren 1999:35) Large schools may have been an administrative necessity, or so it was argued, but it was also an ap- ple of contention. The idea of ‘Storskolan’ (The Large School), as it was called at the time, was heavily criticised. Impersonal, factory-like conditions were not suitable learning environments for young pupils. To those who held this view, the idea that large schools produced mobbing had an in- tuitive appeal.

19 tional credibility and urgency by being linked to other recognised ‘problems’.24 This peculiar phenomenon, which we will refer to as issue-symbiosis, contributed further to the reception and entrenchment of the concept. As more people and organisations – alerted and alarmed by the articles in Da- gens Nyheter – became involved in the discourse, mobbing began to interlock with a host of other issues. The concept of mobbing proved very susceptible to contagion: if a problem was ‘in the air’, it was likely to be linked up with mobbing. In some cases, the opposing camps of a contemporary debate attempted to hijack ‘mobbing’. This was the case in the debate on school organisation and school democracy. According to some observers, the causes of mobbing were to be found in the use of group exercises in school, the lack of discipline, or expe- riments with school democracy.25 Their progressive opponents came to the opposite conclusion, as in the following report from a meeting organised by the Immigrant Women’s Association and the Red Cross. Rigmor von Euler, head of the school welfare officers in Stockholm, was one of the speakers:

She [Euler] reported the following observations: • Large schools: More mobbing • Large school classes: More mobbing • Authoritarian teachers: More mobbing • Democratic schools: Less mobbing This was followed by a round of applause from the crowded lecture hall, to which mainly staff involved in pupil welfare functions had come.26

Euler’s observations were, needless to say, personal impressions rather than established correlations. The first two draws on the interview with Heinemann in Dagens Nyheter, while the latter are Euler’s own suggestions. All the purported

––––––––– 24 Mobbing was often attributed to school size in the news. See for example Dagens Nyheter 13 No- vember 1970; Svenska Dagbladet 12 February 1972; Svenska Dagbladet 21 September 1972; 25 Cf. Östgöta Correspondenten, 16 December 1970. 26 Dagens Nyheter 13 November 1970. The meeting was held in ABF’s premises, and was chaired by the Social Democratic Minister of Education Ingvar Carlsson. Other speakers on the platform were Inga Kempe (chairman of Save the Children Sweden), Peter-Paul Heinemann (who introduced the concept, but who is here listed as a parent representative), Margareta Strömstedt (representing Swedish Radio and Television), Olle Österling (Head of Division at the Board of Education), Lars Hildeberg (PTA representative), Karl Grunewald, (member of the Royal Medical Board), Birgitta Corrias (Head of Division, Swedish Migration Board), Divna Nedelkovic (a schoolgirl who had arrived in Sweden from Yugoslavia three years ago).

20 correlations tie mobbing into a cluster of contemporary issues, thereby making issue-symbioses possible. The early phase in the career of the concept of mobbing contains several examples of conceptual contagion, where ‘mobbing’ absorbed surrounding news or current debates. A complete review of them would take us too far afield. But there is one particularly important area that cannot be omitted. It concerns the relation between mobbing, physical deviance, and minority groups. Peter-Paul Hei- nemann had been very clear that it was the physically deviant children who became victims of mobbing. Moreover, the context in which he placed mobbing (apartheid), and the examples he offered (often recounting his son’s experiences), firmly linked mobbing with race, adoption, and immigration. These issues were highly conspicuous at the time. Adoption of foreign-born children had multiplied during the 1960s.27 So had migration. Between 1963 and 1965, the influx of migrants doubled, primarily due to labour migration. At- tempts to regulate immigration in 1967 temporarily brought down the figures, but the migration flow rapidly recovered and reached an unprecedented peak in 1970 (Nilsson, 2004:19; cf. also Frank, 2005). At the time when the word ‘mobbing’ reached the general public, adoptee children from Korea and different parts of Africa – and migrants from Finland, Yugoslavia, Italy, and Greece – were numerous enough to be noted and few enough to appear exotic in Sweden. Thus, when mobbing was discussed in the press, the victims in the stories were often adoptee children or children of emigrants. Gypsies, particularly, were frequently mentioned as victims of mobbing. Gyp- sies had no doubt long suffered from prejudice and hostility in Sweden, but the reason they were singled out in stories of mobbing was most likely the sudden publicity they received in 1969: this was the year when Katarina Taikon’s immensely popular children’s book Katitzi appeared, and several cases of deporta-

––––––––– 27 This was due to a shortage of native-born children available for adoption, leading among other things to the formation in 1969 of two non-governmental organisations for parents who had adopted, or who wished to adopt, children from other countries. See Adoption – till vilket pris?, SOU 2003:49, pp.83f.

21 tions of gypsy families caused outrage in the press.28 Those who read the series of articles on mobbing only had to turn the page to find reviews of Katitzi or read about the inhumane treatment of asylum seeking gypsy families. The story told in the following letter-to-the-editor, published only ten days after the first article on mobbing in Dagens Nyheter, testifies both to the rapid diffusion of the concept and to the operation of a mechanism of conceptual contagion:

The other day, our ten-year old daughter told me that her teacher had talked to the class about mobbing. That is good, I said. And she had said something like this: Our Swedish children are often very mean to Negroes, Gypsies and Korean children. Only, my daughter added, there is no mobbing in our class. But surely, I said, mobbing doesn’t only affect Negroes and Korean children. There are lots of children who get mobbed because they are different from the others, and whom we also should feel sorry for: children who dress differently, who wear glasses, who are not as tough as the others. Our daughter then told us that the teacher had asked the children how you could see that someone was a gypsy (!). You can see it on their clothes, one boy answered. How can you see it, the teacher asked (still according to my daughter, of course). Well, another kid said, they have earrings. Upon hearing this, one kid said: Britta wears earrings, but she is not a gypsy! The result of it all was that Britta is now called gypsy in the schoolyard during the breaks. This surely cannot have been the in- tention.29

Equally important, from a diffusion and entrenchment point of view, immigrant organisations actively participated in putting mobbing on the agenda, and state officials responsible for immigration issues figured prominently in discussions about mobbing.30 The link between ‘mobbing’ and ‘immigrant’ issues thus had organisational and institutional backing. But mobbing was not equal to ––––––––– 28 Katitzi is the story of a gypsy girl, and of the appalling prejudice she encounters. The author Katarina Taikon had in fact become something of a celebrity already before the publication. She had appeared in the radio show Familjespegeln, and was thereby much encouraged to write the book (cf. the interview with Karin Wilhelmsson in Vår grundade mening 10 August 2006). Tai- kon’s Katitzi has continued to be associated with mobbing. It is listed as suggested further reading in the appendix to the Swedish edition of Erling Roland’s book on mobbing, (Roland, 1983:116). 29 Letter-to-the-editor from Monica Fuchs in Lidingö, published in Svenska Dagbladet 23 November 1969. 30 As mentioned in note 26 above, the head of division of the Swedish Migration Board was on the platform at the hearing. In 1972, Kaisa Langton, also representing the Swedish Migration Board, wrote an article in the official journal of the National PTA in Sweden – Barn i hem skola och sam-

22 racism: it was the release of aggressive energy stored up in groups of understimulated children, unleashed upon those who were ‘physically deviant’ from their peers. Immigrant and adoptee children formed an important subset of that class, but they did not exhaust it. Organisations representing other ‘physically deviant’ groups also responded. Less than three weeks after the first article in Dagens Nyheter, the Swedish Disability Council – representing eighteen separate organisations, and ten other associations – petitioned the government about mobbing.31 Whether the ‘deviants’ were handicapped or immigrants, mobbing was intertwined and occasionally conflated with minority-group issues, as illustrated by the brief newsflash below:

Mobbing. The government has turned down a request from the National Parent-Teacher Association and Seco [a national student organisation] for a sum of 35.000 Swedish crowns, for a prospected information campaign about minority-group problems (mobbing) in the schools.32

More research is no doubt needed on how ‘mobbing’ was taken up by different non-governmental organisations, how they interacted with each other and with state agencies, and how this affected the diffusion and entrenchment of the concept. The available evidence so far suggests that the process can be understood as ‘institutional bootstrapping’: the launch of the concept in the media triggered activity in a variety of interest organisations, whose members and officials amplifi- ed the message, emphasised those parts of it that fitted the member group, and contributed examples and amendments that fed back into media, literature, politics, other organisations, and an amorphous collective consciousness. At the same time, and as the next example will show, organisations were not the only agents to make amendments to the concept. By the end of February 1970, less than four months after the first article in Dagens Nyheter, it was opening night for the first stage play about mobbing.33

hälle – where she argues that immigrant children are particularly likely to become victims of mobbing. There are counter-examples: see Dagens Nyheter 22 February 1970. 31 Dagens Nyheter 3 December 1969. 32 Svenska Dagbladet 18 December 1970. 33 Reported in Dagens Nyheter (Nordost) 5 February 1970. Novelists were equally quick in their response. According to one observer, mobbing was the dominant theme in children’s books in 1970.

23 Its plot contains some elements deriving from Peter-Paul Heinemann’s conception of mobbing and some that are of the playwright’s own device. The chief victim in the play is a boy who has moved from Skåne to Stockholm and speaks with a broad accent. A second victim is overweight. Both are ‘deviant’. They have physical characteristics that set them apart from the other boys.34 This was integral to Heinemann’s view of mobbing. But Sjöstrand’s play also builds up a conclusion that is not found in Heine- mann: ‘On the deepest level, mobbing occurs because the perpetrators are insecure; afraid of falling victim of mobbing themselves, they are prone to mob others as a means of gaining acceptance and esteem from their peers’.35 Sjö- strand’s play may well be the immediate source of this still widespread piece of folk psychology about the nature and causes of mobbing, but this idea too enveloped notions coming from elsewhere. The playwright had sought expert advice: academic psychologist Jan Fröberg had been enrolled to assist her in writing the play.36 We will have more to say about the role of science, but before we get there it is useful to provide a model of the sequence of events.

Cf. Ying Toijer-Nilsson, (1970) ”Mobbing kallas det”, Barn i hem skola och samhälle, 24, h.9, s.33-35. Literary portrayals of mobbing were clearly in demand. 23 November 1969, librarian Anne-Marie Alfvén-Eriksson was interviewed in Dagens Nyheter about mobbing. She was worried that it was difficult to find good children’s books about mobbing. In one sense, this is hardly sur- prising, considering that the word had only been launched ten days ago. But this is only partially true. Once the concept was out there, old works of fiction could be re- categorised under the new label. These works would then become part of the available stock of examples that – in the absence of a denotative definition – defined mobbing. Manuals, educational materials, and scholarly publications relating to mobbing often contain lists of recommended works of fiction. It would be a research project in itself to charter how this ‘literary canon’ has evolved. 34 This is further underlined by Sjöstrand in the preface to the published version of the text, where she instructs those who want to stage the play that it is not necessary for the victim to have a south- Swedish accent: he could speak in any deviant idiom, or he could speak like the rest of the actors provided that his status as deviant is marked out in some other way, e.g. by referring to him as a Yugoslav or Finn (Sjöstrand, 1973: 6). 35 This understanding of the ultimate causes of mobbing is signalled already in the title of the play, which roughly translates as Are you afraid of Uffe? He is also afraid. The explanation was also brought up by the playwright on national radio – in Familjespegeln 14 March 1970 – and her explanation was granted to be ‘most probable’ by commentators in the press (Svenska Dagbladet 15 March 1970). 36 Cf. Svenska Dagbladet 15 March 1970. The practice of consulting psychologists on matters of mobbing is found elsewhere. Swedish television employed a team of psychologists to evaluate two prospected television programs about mobbing. See Dagens Nyheter 13 November 1970. .

24 5. The early image of mobbing: an analytical model

Let us recapitulate the argument so far. The absence of a strict denotative definition, and the associated lack of systematic empirical knowledge about the phenomenon, in conjunction with a massive public response, created a situation where those who contributed examples and accounts filled in the picture as they went along. In this filling-in process, which characterised the discourse on mobbing in the years immediately after 1969, contributors drew on two types of sources. First, there was the stock of already available examples and accounts, particularly those furnished by Peter-Paul Heinemann. New examples of mobbing were grafted upon or analogous to previous ones. Available conceptions and accounts of mobbing were passed on as quotes embedded in quotes. This is evident in the examples given above: Heinemann’s impression, that large schools and large school classes are conducive to mobbing, was repeated by Rigmor von Euler at the hearing organised by the Immigrant Women’s Association and the Swedish Red Cross; Ingrid Sjöstrand’s play reiterated Heinemann’s idea that mobbing af- fects those who are physically deviant, others forwarded the notion that immigrant or adoptee children become victims of mobbing, etc. Heinemann’s message was spoken with many voices, thereby creating an impression of indepen- dent corroboration. The second type of source to draw upon was more heterogeneous. Those who wrote and talked about mobbing contributed their own understandings of the nature of mobbing, on the basis of their own experiences, textbook wisdom, or a general idea of what was likely to be its causes. Rigmor von Euler added authoritarian teachers and lack of school democracy to the list of likely causes; Ingrid Sjöstrand held the insecurity of the perpetrators to be responsible. To the extent that such additions were passed on, they too became part of the available stock of examples and accounts of mobbing.

25 Although the latter class of relatively unfettered contributions did contain idiosyncrasies, heterogeneity was not infinite. They too were patterned, albeit by different mechanisms. One such mechanism is what we may call conceptual contagion: the infant concept absorbed current issues. This was the case already in Heinemann’s conception where mobbing was linked up with urbanisation, school size, immigration, and adoption. Each of these issues was on the agenda in Sweden by the late 1960s and early 1970s. Other additions also smacked of Zeit- geist, e.g. when the materialist style of life in Western society was pegged down as the cause of mobbing (Nielsén & Stigendal, 1973). Conceptual contagion was especially likely to catch on, and fuse with the meaning and understanding of mobbing, if it was combined with one or more sup- porting mechanisms: (a) issue-symbiosis, (b) institutional backing, and (c) proximity to the stock of available examples, particularly those relating the phenomenon to ethology. If mobbing was causally linked to other recognised problems, be it urbanisation or immigration, the new concept would resonate with those who were concerned about these issues as well. To the extent that the sup- posed causal links alarmed organised groups and state agencies, issue-symbiosis received institutional backing. And the causal link appeared all the more credible if it had already been reported by a plurality of voices and was thought to be founded upon hard science ethology. These mechanisms operated jointly to produce the meaning of the concept ‘mobbing’ in the period between 1969 and 1973. The publication of Dan Olwe- us’ book on mobbing in 1973 would prove to be a watershed. Until then, the build-up of examples and accounts had remained unchecked by scientific investigation, save for the loose inferences from ethology. This changed with Olweus. His research challenged beliefs that had become entrenched and taken for granted and contributed to altering the very meaning of the concept. As we shall see, however, Olweus’ research was bound up in intricate ways with the way in which the lay concept had developed.

26 6. Mobbing and the scientific literature

Peter-Paul Heinemann’s original conception of mobbing, as presented in Dagens Nyheter, enveloped and built upon Konrad Lorenz’s work on aggression in animals. Mobbing was a subtype of aggression, where a group of individuals teamed up against a single individual. It was, again in consonance with Konrad Lorenz’s views, the instantaneous release of stored up aggressive energy. This way of framing the phenomenon loaded the dice in favour of some scientific disciplines over others. Aggression was the domain of ethology and psychology. The fact that mobbing was mainly associated with school settings also served as an invitation to scholars in education departments, and all the more so since educational psychology was a dominant strand in the discipline at the time. This is not to say that other disciplines, e.g. sociology, took no notice of the phenomenon or that their voices are absent from the discourse. Mobbing was the topic of several student seminars in sociology at Stockholm University between 1970 and 1973 (Nielsén & Stigendal, 1973:7). And Dagens Nyheter had in fact interviewed Kerstin Elmhorn, a sociologist specialising in juvenile delinquency, for the series of articles about mobbing. But with mobbing couched in terms of aggression, sociologists were somewhat disoriented and had to relate as best they could to the ethological imagery. This is what Kerstin Elmhorn did in the interview:

- It is good that we have this new concept. The advantage of using the expression mobbing is that the biological explanation is also sociologically correct, i.e. you mob because you are afraid of deviant and weak individuals who can threaten the survival of the species.37

A booklet published in 1973, by one student and one senior sociologist in Stock- holm, provides a similar example: while their aim was to lay the blame for mobbing on Western individualism and materialism, the first chapter was neverthe-

27 less on aggression. In sum, then, aggression ill suited sociologists as theoretical category, and their contributions to the literature have been sparse and far between.38 Instead, it was psychology and education scholars who took a scientific interest in the phenomenon. Two early contributors stand out, Dan Olweus (psychology) and Anatol Pikas (education). To this list we should add psychologist Heinz Leymann. He transferred the concept of mobbing to adults and to bullying in the workplace and it was via his writings that the word mobbing entered the German language. But Leymann wrote in the 1980s, and a lot had happened to the concept in the meantime. Mobbing had already entered Norwegian and Da- nish vocabulary, albeit by a different route. First Heinemann and then Olweus had an attentive audience in these countries in the first half of the 1970s, and particularly in Olweus’ version mobbing was limited to childhood and to school settings. Even more importantly, the meaning and understanding of the concept had undergone changes as a result of its reception in the sciences. Dan Olweus Hackkycklingar och översittare: Forskning om skolmobbning (1973), which roughly translates as Whipping Boys and Bullies: Research on School Mobbing, was the first proper scholarly publication on mobbing. There are some attempts to study mobbing empirically that antedate or parallel his endeavour. (The School Board in Stockholm had commissioned inquiries, and there were quite a few student papers on mobbing in the early 1970s.39) None of these compare to Olweus’ research in terms of scope and ambition. With the support of large scale, quantitative investigations, he set out to question established opinion about mobbing. But as we shall see, Olweus’ research was curiously bound up with the lay views he challenged. The following three sections deal, in turn, with how Olweus related to the notion of mobbing cum aggression, with

37 Dagens Nyheter 23 November 1969. 38 The next sociological publication on the topic, after Nielsén & Stigendal (1973), was Roos (1979). After that we have to wait until 2001 for the next research contribution (Eriksson, 2001), although mobbing was included as a topic in a student textbook on social psychology (Angelöw & Jonsson 1990). 39 See for instance Abrahamsson et al, 1973; Andersson, 1974; Brandt et al, 1972; Christensson, 1974; Granath et al., 1973; Gällingsjö & Hallmo, 1971; Hansson & Ödman, 1973; Hillner, 1974; Holmberg, 1974; Jeppson et al 1973; Larsson, 1974; Nilsson 1974; Norrman, 1974; Sjölander, 1973; Wramner & Åhlander, 1973.

28 how the meaning of mobbing changed, and with Olweus’ empirical results and theoretical framework.

6.1. Dan Olweus and mobbing cum aggression Konrad Lorenz’s work on aggression in animals suggested that mobbing was a subtype of aggression. This was an open invitation to psychologists, but it had special significance for Dan Olweus. His doctoral dissertation, published in 1969, was an inquiry into the psychology of aggression. Its contribution was as much methodological as it was theoretical, being concerned both with the intri- cacies of projective testing and with understanding the mechanisms of aggressive behaviour. In the author’s own formulation, it was an attempt to ‘predict overt aggression in an interpersonal situation on the basis of aggressive responses to a specially constructed projective test’ (Olweus, 1969:11). Thus when mobbing emerged on the public agenda, Olweus was already equipped with psychological theories and methods suitable for the study of aggression. This no doubt helps to explain how it was possible for him to initiate, conduct, write up, and publish his research in the short time span between November 1969 and 1973. If mobbing was aggression, then Olweus could transport his previous findings and methods to what everyone was now talking about. There are several traces of Olweus’ doctoral thesis in his later work on mobbing. The subjects in the thesis are twelve to fourteen year old school boys; in the later book they are twelve to sixteen year old school boys (Olweus, 1969: 27; Olweus, 1973:9). Projective tests, the main concern in the thesis, were used in the research project on mobbing but were sparingly reported in the book (Olwe- us, 1973:36,142). Other methodological devices are directly passed on from the first to the second book. The dependent variable of the doctoral thesis was ‘overt aggression’, measured as ‘Tendency to start fights’. Four subjects, randomly drawn from each school class, were given a set of white cards with the names of the boys in the class printed on them. They were then presented with a slip of paper, headed ‘He starts fights’, and were asked to rate each of the boys in the class on a 7-point scale, ranging from ‘very seldom’ to ‘very often’, by placing

29 the name card on the paper slip (1969:16,35). The same variable and the same procedure are used in Olweus’ investigation of mobbing, and the rating procedure is used on other items as well, such as ’Tendency to be the target of other people’s aggression’, etc. (Olweus, 1973:41f). We will discuss Olweus’ variables, and the underlying model of mobbing they are part of, in more detail below. For the time being, it suffices to note that the Lorenzian conception of mobbing as a subtype of aggression made it possible for him to transfer his research on aggression to the new issue of mobbing. Yet the centrality of Konrad Lorenz in the early discourse on mobbing was as much a curse as it was a boon. Olweus may have benefited from Konrad Lorenz’s notion that mobbing was best conceptualised as a form of aggression, but he did not share the ethologist’s view of aggression. In particular, he was critical of Lorenz’s energy model of aggression, which he found to be vague, muddled, and incorrect. ‘The conception of a biological influence in the form of constantly active aggressive drive’, he concludes, ‘should be rejected for a number of reasons’ (Olweus, 1970:6). He sides instead with reactive theories of aggression and embraces the Yale group’s frustration-aggression model, if not to the point of wholesale acceptance:

In my opinion, both situational factors and the individual’s more stable reaction tendencies should be taken into systematic account in an adequate theory of aggressive behaviour (Olweus, 1970:8).

What Olweus tried to do in his thesis and in related papers was to construct and test just such a model of the mechanisms behind aggressive behaviour. Stable personality traits – habitual aggressive and aggression inhibitory tendencies – have a key role to play in that framework. These tendencies, he continues, func- tion as ‘dispositions both to appraise certain stimulus situations in a particular manner and to respond to such appraisal with relatively consistent reactions (at a certain strength)’ (Olweus, 1970:10). This trait approach was later transplanted to the study of mobbing (Olweus, 1973:23). Konrad Lorenz concept of mobbing paved the way for Olweus’ research, but ‘mobbing’ carried with it an unacceptable view of aggression. Olweus in fact had doubts about the appropriateness of using the word – not only because of its proximity to Lorenz but also because of the inherent vagueness and ambiguity of

30 a concept that had evolved freely in everyday usage and accumulated meanings in the process (Pikas, 1975:125). It is no coincidence that ‘mobbing’ occurs only in the subtitle of his 1973 book on mobbing: he preferred to conceive of the problem in terms of ‘bullies’ and ‘whipping boys’ (Olweus, 1973:9). But however much Olweus may have hesitated, the book was nevertheless said to be about school mobbing. So instead of relinquishing the concept, he reformed it so as to be compatible with his own understanding of aggression. This meant, among other things, a subtle break with the original understanding of mobbing as portrayed by Peter-Paul Heinemann. The metaphor of the deviant child acting as a lightning conductor makes no sense outside the framework provided by an energy model of aggression. It was, moreover, the idea of a specific aggressive energy continually building up in the organism that allowed Heinemann to extend the ethological argument and say that city size, school size, and class size were instrumental in producing mobbing: metropolis and suburbia do not provide acceptable outlets for children’s mounting aggressive energy, and mobbing is the result of their active search for release. That syllogism loses credibility if the energy model of aggression is abandoned. This difference was no doubt too subtle to be perceived by the public. Heine- mann and Olweus were sometimes quoted as basically saying the same thing, the two being lined up abreast in the combat against mobbing.40 Olweus’ concept of mobbing, not to mention his findings and conclusions, set him apart from entrenched conceptions in other respects as well, but two factors worked to obscure the breech, at least in the public eye. First, the uses and understandings of the concept of ‘mobbing’, as they unfolded in the discourse, were far from unitary. Once Heinemann had asserted that large cities and large schools produced mobbing, the explanation was freely dis- posable without reference to Konrad Lorenz. That reference facilitated the reception. Yet the reason the assertion caught on had more to do with issue-symbiosis than with syllogisms and inferences. Lorenz remained a central source as the discourse evolved, but the usages that cropped up were not necessarily compatible

––––––––– 40 See, for example, Motion 1974:724 in the Swedish Riksdag.

31 with his discussion of mobbing. Peter-Paul Heinemann himself was no exception:

Many people think that mobbing necessarily means that a group of children beats up a peer whom it has chosen as victim, says Peter-Paul Heine- mann, who in the end of June gave a lecture to school staff during a course on mobbing. But there is a form of mobbing that is both immensely crueller and more difficult for adults to detect. It consists in freezing a person out. Suddenly, one of the children in a group is treated as if he was an abject thing rather than a human being.41

In this newspaper interview, Heinemann extends the class of mobbing behaviours to encompass exclusion, the ‘freezing out’ and isolation of a peer. He was not the only one, or the first, to do so. Nor was he the last: exclusion in this sense is still considered to be a key instance of mobbing.42 By doing so, however, the relation to Lorenz’s mobbing birds becomes strained. The parallel with a flock of grazing birds attacking a predator is plain enough as long as we conjure up the image of a bunch of kids beating up a single child. Not so with a group of children shunning a singled out peer. And prolonged collective avoidance does not quite fit the imagery of mounting aggressive energy explosively released upon a hapless victim. It does align, however, with Heinemann’s contention that mobbing is a situation of ‘all against one’. The meaning of mobbing evolved organi- cally, with contributions from many authors, and they were not all consistent with the ethological model of aggression. Olweus’ departure from it was not uni- que and could easily go unnoticed. Second, Dan Olweus used a terminology that generated a mirage of continuity with ethology. The Swedish word ’hackkyckling’ in the title of Olweus’ book on mobbing was translated above as ‘whipping boy’. This is the formulation Olweus used in the English translation and in some of his Anglophone publications. But the rendering does not quite capture the connotations of the word. The context to which it belongs is the work of the Norwegian zoologist Thorleif

––––––––– 41 Göteborgs Handels- och Sjöfartstidning 13 July 1971. 42 The notion of ’shunning’ has even generated a small literature of its own, see Tanaka (2001).

32 Schjelderup-Ebbe on the ‘social psychology of domestic chicken’.43 From his writings came the familiar concept of pecking order, Hackordnung in the Ger- man original (hacken = ‘to peck’).44 This was not the only conceptual innovation in Schjelderup-Ebbe’s articles on social dominance hierarchies among groups of fowl. Alongside Hackordnung we find a series of compound nouns involving hacken – e.g. Hackliste, Hackkombination, Hackrichtung, Hackzustand, and Hackgesetz – none of which received commonly understandable English equiva- lents. The term ‘Hackkyckling’ in the title of Olweus’ book was introduced with reference to Schjelderup-Ebbe (Olweus, 1973:19). Curiously enough, Hackhuhn – which would have been the German word for ‘hackkyckling’ – was never used in these articles. The word nevertheless had a familiar ring: it had found its way into Swedish dictionaries long before Schjelderup-Ebbe used it.45 While Schjelderup-Ebbe dealt in abstract relations, the vulgate Hackhuhn is a familiar, concrete thing; but rather than relying on the vernacular sense, Olweus explicitly refers to his articles.46 And it was not difficult to see what Hackhuhn could mean within Schjelderup-Ebbe’s framework: ––––––––– 43 Cf. Schjelderup-Ebbe (1922; see also 1913; 1921; 1924). The idea of a social psychology of fowl may seem quaint to a modern reader, but prior to World War II attempts to fashion an ‘animal sociology’ were not uncommon. There is a long list of works along these lines – from Alfred Espinas Des sociétés animales (1887), over Karl Escherich’s Termitenleben auf Ceylon: neue Studien zur Soziologie der Tiere zugleich ein Kapitel kolonialer Forstentomologie (1911), to Friedrich Al- verdes’ Tiersoziologie (1925). The word ‘peck order’ entered the English vocabulary through a translation of the latter work, published in 1927 (Perrin 1955). Schjelderup-Ebbe himself was ap- pointed professor of sociology at Université nouvelle de Paris in 1931. Animal sociology did not quite perish with the war: the 1952 edition of Encyclopedia Britannica had an article on animal sociology, and Schjelderup-Ebbe’s 1924 article on the social psychology of birds was reproduced in Carl Murchison’s Handbook of Social Psychology (1949). 44 As Porter Perrin observed, zoologists have tended to translate Hackordnung as ‘peck order’, while social psychologists prefer ‘pecking order’ (Perrin 1955:265). 45 The first use of the term recorded in Svenska Akademiens Ordbok dates from 1813. 46 Note that Olweus’ choice of term had a double advantage. The reference to ethology preserved a somewhat spurious sense of continuity with Konrad Lorenz. But the word was also a widespread commonsense term, which had been used to name approximately the same phenomenon described by mobbing. Savour the following description – in Statens ljud- och bildarkiv – of the television program Skol-TV: Vad tycker du? 2. Särlingen: A classroom with 10-12-year-old children. A girl and a boy are identified as deviant. Images from a pultry farm with a pecking-chicken [Hackkyckling]. Images of monkeys at Skansen Zoo. Images of vulnerable children in a schoolyard. This was broadcasted on 22 November 1967, i.e. two years before mobbing entered Swedish vocabulary. In other words, Olweus’ choice of terms also evoked familiar lay imagery.

33 literally meaning ‘pecking chicken’, it denoted the chicken at the extreme bottom of a pecking order, which is pecked without pecking in return. By thus invoking the study of animal behaviour, Olweus’ book was likely to be perceived as a continuation of the ethological Denkfigur found in Konrad Lo- renz. The proximity, however, is illusory. Despite the title of the book – and despite the reference to Schjelderup-Ebbe and a few other works on animal behaviour – Dan Olweus was sceptical of inferences from animal to human behaviour. His final verdict was that the study of animal behaviour may be a useful source of ideas, but that we must be very cautious about drawing parallels between animal and man (Olweus, 1973:20f). And as we have already seen, Olweus was critical of Konrad Lorenz’s theory of aggression. The concept of mobbing, on the other hand, might be salvageable provided that it could be rid of the Lorenzian view of aggression it was associated with. This is exactly what Olweus attempted to do in his 1973 book on mobbing: he had formulated his own, psychological model of aggression that he wished to apply to the newly discovered phenomenon. The results of his investigations contradicted just about every single element in the image of mobbing that had become entrenched by then. His way of addressing mobbing also pushed him towards a specification of the concept that removed it from its original sense.

6.2. The meaning of ‘mobbing’ Anatol Pikas, in his first book on the topic, complained that Dan Olweus’ book did not provide a definition of mobbing (Pikas, 1975:14). He had defined ‘whipping boy’ and ‘bully’, but had left the key term undefined. Later on, Olweus would supply a definition of mobbing that has become standard in the literature. Mobbing, according to this definition, is when ‘one or more individuals are subjected to negative actions, on several occasions and over an extended period of time, by one or more individuals’ (Olweus 1986:8). ‘Negative actions’, in turn, are understood as ‘when an individual inflicts, or tries to inflict, harm or discom- fort upon another individual.’ There are a host of parallel definitions of ‘bully-

34 ing’ and ‘mobbing’ in the scholarly literature, but this one is adhered to, or at least related, in an enormous amount of articles. This is, moreover, the definition found nowadays in Swedish official school documents.47 In 1973, Olweus did not offer this formula as the defining characteristics of mobbing. But some of its central elements do appear in the book, albeit in another guise:

In the final pages of the previous chapter I underlined that my focus of interest in these research projects has been with whipping-boy-problems of a certain duration, i.e. where the problem has been observed several times and for an extended period of time. The main reason for this choice is that such problems – if they exist – would appear to be the most serious: they are the ones likely to cause most suffering and they are liable to engender the most injurious future consequences for the affected individual. (Olwe- us, 1973:18; italics added)

The specification that the victim must have been subjected to negative actions several times and for an extended period of time is presented here as a particularly important subset of the undefined superset mobbing. Once it had been formulated it could begin to slide, and the formula was soon promoted to the status of definiens of the superset. Anatol Pikas, who chided Olweus for not defining ‘mobbing’, in fact echoed some of the relevant parts of his formulation in his own proposed definition of mobbing.

Mobbing, as I intend to use the term, refers to repeated negative activities (physical and mental assaults and/or exclusions from the group) directed against a single individual by two or more interacting individuals. (Pikas, 1975:16)48

––––––––– 47 We find it, for example, in the action-plans against mobbing, which every school is obliged to produce. A small sample of schools suffices as demonstration: Strandängsskolan in Båstad, Rydskolan in Skövde, Bosvedjeskolan in Sundsvall, Ungärde skola in Malung, Sunnadal och Holmsjö rektors- område Karlskrona, Tunabergsskolan in Uppsala, Söraskolan in Österåker, Rönnängs skola in Tjörn, Oscarsgymnasiet in Oskarshamn, Vilboksskolan in Oskarshamn, Ravebergsskolan in Göte- borg, Romaskolan in Gotland, Teleborg Centrum in Växjö, etc. 48 Pikas added a couple of other specifications to his definition. The ’activitites’ must be illegal, for example, so as to set mobbing apart from cases where several police officers chase a criminal.

35 The major difference between Olweus and Pikas is that the latter reserves the term mobbing for situations where there are two or more interacting perpetrators. This brings us back to the meaning of mobbing in ordinary language, as it evolved prior to the publication of Olweus’ book, for Pikas motivates the restriction with recourse to how the term is used in everyday speech. ‘It appears that practi- cally everyone who has written about mobbing’, he notes, ‘agree that mobbing involves two or more perpetrators’ (Pikas, 1975:15). The locution ‘A mobs B’, he continues, can only be the product of an unreflective use of language. Were we to accept it, mobbing could just as well be replaced by ‘aggression’, ‘fight’, or ‘conflict’. Pikas’ remarks make sense, considering how the discourse on mobbing began and evolved. A mob must surely consist of more than one person. In ethology, mobbing animals were entire flocks, and Peter-Paul Heinemann described mobbing as situations of ‘all against one’. Mobbing was group violence. But linguistic agreement was less unanimous than Pikas lets on. In Dan Olweus’ conceptua- lisation of the phenomenon it was perfectly possible to speak of mobbing when there was just one perpetrator. The notion of mobbing as ‘group violence’ poorly fitted Olweus’ theoretical framework, where aggression was perceived as the result mainly of stable, individual reaction tendencies. Here, too, the concept was in need of reform, and once again the conceptual shift was quite subtle. A book on this topic could not ignore Peter-Paul Heinemann’s assertion that mobbing is a situation of ‘all against one’. But was it necessary that each member of a group had an equal share in the activity? This conclusion would be to press the concept too far, considering the variety of ways in which it had already been used. Olwe- us exploits this possibility.

If situations of the type ’all against one’ emerge in the school environment – and they probably do from time to time – it is nevertheless likely that different members of the group take unequal part in the activity. It is, above all, probable that there are one or two active initiators, or instigators, who are instrumental in bringing the situation about in the first place. By conceiving of mobbing as aggression from a unitary, rather anonymous group, we may all too easily lose sight of the active role of ‘bullies’. (Ol- weus, 1973:13)

36 By admitting that all-against-one situations do occur, Olweus provides a sense of continuity with Heinemann and entrenched discourse. But the gist of the para- graph is that such situations do not exhaust the class, and that focussing on them tends to obscure the probable causes of mobbing. In fact, he goes on to say, isn’t it likely that situations of this type are but a tiny fraction of this (still undefined but no doubt enlarged) class of mobbing behaviours? Olweus is inching his way from the Heinemann-Lorenz concept of mobbing and towards a conception more compatible with the trait approach used in his research on aggression:

We often find subgroups – cliques – within a school class and an aggressive subgroup could probably oppress and torment other children rather intensively and on a large scale. Such phenomena are easily overlooked if the collective aspects of mobbing are given undue emphasis. If this type of ’mobbing’ – by very small groups or even by single individuals – is the more common in our school environments, then the current use of the concept can make it difficult for teachers to detect the phenomena in their immediate environment. (Olweus, 1973:14; italics added)

What is presented as an empirical argument is in fact a conceptual shift of ground. Mobbing is no longer group violence. As for Pikas’ charge that the notion of single-perpetrator-mobbing equates ‘mobbing’ with ‘aggression’, Olweus would have had no problem in pleading guilty. To him, mobbing was a special case of aggression, and this much was perfectly consonant with Konrad Lorenz’s and Peter-Paul Heinemann’s usage.49 In Pikas’ view, Olweus’ reformation of the concept was an abuse of language, and he has stuck with the view of mobbing as group violence (cf. Pikas, 1998:14f). Yet Pikas went along with Olweus attempted reformulations on one important point. As we saw above, he incorporated into his own definition the requirement ––––––––– 49 The following quotation from Olweus simultaneously captures his reservations regarding the term ‘mobbing’, reiterates his reasons for keeping it, and shows that he is – nevertheless – mainly preoc- cupied with the psychology of aggression: I call this book Whipping boys and bullies, which … says something about my view of the problem. I have nevertheless chosen to keep the term mobbing – which is also included in the subtitle – since this term is well known to the general public. In the interest of clarity, it is perhaps necessary to underline the principal point of departure of the researches presented here: to study the incidence of and mechanisms behind different forms of aggression and op- pression directed at single children/youth by other children/youth during an extended period of time. (Olweus, 1973:16)

37 that the victim is subjected to negative actions several times and for an extended period of time. We have grown so accustomed to seeing this as a defining characteristic of mobbing that it is easy to overlook that it removes some core examples from the set of behaviours classified as mobbing. Recall, for example, how Heinemann described mobbing in the first article in Dagens Nyheter. He conjures up the image of a group of understimulated city children waiting around for something to happen. With no suitable outlet for their constantly mounting aggressive energy they will seize the first opportunity to unleash their aggression. ‘When you pass through such a crowd of kids with a Negro or Korean child by your hand’, Heinemann says, ‘they immediately start to mob’.50 Now, once repetition and duration had been introduced as definitional criteria, such isolated incidents of instantaneous, spontaneous releases of aggressive energy were no longer examples of mobbing. In Olweus’ 1973 book, repetition and duration are not offered as part of a definition of mobbing. These specifications are introduced much the same way as he introduced the notion of single-perpetrator mobbing, i.e. by arguing that mobbing of this type merits special attention (Olweus, 1973:14f). He agrees that there probably are cases where a capricious crowd acts in the spur of the moment and pours its aggression and hostility over a child who is unknown to them. And there is no doubt, he added, that this must be most unpleasant to the victim. By un- duly emphasising such cases, however, we are liable to lose sight of the more important set of behaviours that is the focus of Olweus’ own research on whipping boys and bullies. The possibility of single-perpetrator mobbing widens the concept of mobbing; the introduction of repetition and duration as prerequisites narrows it down. Without expressly defining ‘mobbing’, Olweus inserted new conceptual elements that were at odds with previous discourse but in tune with his own research on aggression. This was perfectly clear to at least some commentators at the time:

––––––––– 50 Dagens Nyheter 13 November 1969.

38 Olweus’ questions are entirely focussed on organised mobbing for extended periods of time. He picks out bullies, i.e. boys who ‘rather frequently oppress other boys or girls physically or mentally’. Hence, the possibility of studying collective mobbing disappears. Moreover, Olweus is not inte- rested in transient mobbing problems. Thus, he does not study the occur- rence of spontaneous, collective mobbing. (Nielsén & Stigendal, 1973:64)

The authors claim that Olweus has only chartered one province of what we un- derstand by mobbing, and the reason they can do so is that they have in mind the lay concept that evolved and became entrenched in the period between 1969 and 1973. In spite of Olweus’ efforts, they claim, there was still room for studies of ‘collective’ mobbing (‘all against one’) and of spontaneous (as opposed to durable and repeated) mobbing. Similarly, Anatol Pikas protested against Olweus’ extension of the concept to include single perpetrators. But Pikas also contributed to refining parts of Olweus’ reformulation into a definition of mobbing. Ori- ginally, the proviso that negative activities must occur ‘several times and over an extended period of time’ applied only to Olweus’ concrete research project; in Pikas’ hands, it became part of the definiens. Pikas exemplifies a general pattern. Since Olweus’ 1973 book was the first proper scholarly work on mobbing, subsequent research had to relate to it. By thus quoting Olweus, his framework was – wittingly or unwittingly – encapsula- ted, transmitted, and entrenched. His conception was placed inside those of other scholars, like a small Russian doll inside a larger one. In the mid 1980s, when the discourse on mobbing had a second coming in Sweden, the concept was almost universally defined as a situation where ‘one or more individuals are subjected to negative actions, on several occasions and over an extended period of time, by one or more individuals’. This was Olweus’ definition of mobbing as of 1986. It appears, too, in the works of other Swedish scholars in the same period (von Schéele, 1993:13). As others quoted these (and other, similar) works, the formula was increasingly decoupled from its origin in Olweus and became a free-floating and generally recognised definition of mobbing. Olweus acted as a conceptual entrepreneur; he all but usurped the existing concept, permanently altering its meaning.

39 6.3. Olweus’ empirical results and theoretical framework Olweus’ modifications of the concept of mobbing, consequential though they are, may have escaped the attention of the lay audience. But the results he reported were conspicuous indeed. There was already an accumulated body of purported knowledge about the conditions and mechanisms producing mobbing. These assertions, however, could be confronted head on by empirical inquiry. They could be turned into hypotheses, incorporated in the design of the investigation, and tested against the facts. This is what Olweus did. Among the questions he posed to the empirical material we find some that should be familiar by now:

Is there any truth to the often-made assertion that mobbing increases in direct proportion to school size? Are whipping boys more often found in large school classes? (Olweus, 1973:79)

[P]hysical deviance in the whipping boy … has been accorded decisive importance in the Swedish debate about the mechanisms of mobbing. It is fat children, immigrant children, children who look different or wear unu- sual clothes – for short: deviant children – who become victims of mobbing. Their difference, their deviation from the group pattern, in itself generates tension and irritation in the group, with mobbing as a result. (Ol- weus, 1973:101)

The discourse on mobbing which evolved between 1969 and 1973 had pointed out physical deviance in the victims, school size, and school class size as the causes of mobbing. Dan Olweus had no choice but to include them in his investigation. They were entrenched lay conceptions of the phenomenon, supported by issue-symbiosis and backed up by the involvement of a web of organisations and institutions. His sample, consisting of schoolboys from metropolitan areas, did not allow him to test the equally widespread assertion that mobbing increases in proportion to the size and density of cities. Otherwise, all the major purported causes were included in his 1973 book. On all these counts Olweus’ results contradicted established opinion. ‘Devi- ant’ kids – including immigrant and adoptee children – were not overrepresented

40 in the whipping boy category.51 School size had no effect on mobbing. The size of the school class, and related measures such as the group climate of the class, proved not to be correlated with mobbing. This absence of correlations left the field open for alternative hypotheses, also included in the model (Olweus, 1973:25). This model, then, consists of two components: (a) A core model, drawing on Olweus’ previous research on the psychological mechanisms of aggression. (b) A counter-model, made up of factors that were deemed to be causes of mobbing in the discourse that evolved between 1969 and 1973. Olweus’ inclusion of them was an act of vigilance against spontaneous thinking and common sense. The factors making up the counter-model could, on the face of it, be seen as part of the core model. As we noted above, Olweus’ theory of aggression stated that ‘both situational factors and the individual’s more stable reaction tendencies’ should be taken into account (Olweus, 1973:8). The very fact that the situational factors are gathered from entrenched public discourse testifies, however, that their theoretical importance is played down. In a later publication Olweus asserts that ‘the situation-person issue, as commonly formulated, is probably best regar- ded as intrinsically unanswerable’ (Olweus, 1977:222). He concludes, moreover, that the role of situations may not be as important after all.

Within the aggressive motive area … I have obtained data indicating a considerable degree of consistency across situations, within a particular data source (such as peer ratings) and, more importantly, across different sources of data (e.g., inventory data versus peer ratings of overt aggressive behaviour […]). Very likely, the substantial degree of ‘cross-situational’ consistency obtained was, in part at least, a consequence of the fact that the subjects, 13-year-old-school boys, had known each other for several years and could judge each other’s behaviour with high accuracy. (Olwe- us, 1977: 232)

––––––––– 51 The one exception to this rule was ‘physical strength’: perpetrators tended to be physically stronger, victims physically weaker, than the control group. As Olweus adds, however, this is not was what meant by ‘physical deviance’ in previous discourse (Olweus, 1973:115).

41 So as long as research is done in natural rather than experimental settings, we can safely bracket situational factors and rely on stable personality traits to do the explanatory work. And although this stance may, for all we know, have been the result rather than the premise of his early investigations, Olweus’ core model of aggression and mobbing laid emphasis on stable reaction patterns in the individual. His formulations in the theoretical preliminaries to the 1973 book placed them side by side, without according priority to either situation or person.

A great many research results from the field of aggression (see e.g. Ber- kowitz, 1962; Olweus 1973a) indicate the need to consider, in addition to situational factors, the relatively stable personality equipment of each individual. Here, I am also thinking of such factors as genetic dispositions, parent-child relationships, early experiences, and so on, that may contribute to the creation of relatively stable and typical reaction patterns in the individual (personality factors). (Olweus, 1973:23)

Yet whereas the ‘personality factors’ of the model were derived from Olweus’ own research on aggression, the ‘situational factors’ were basically gleaned from the general debate on mobbing. The empirical analysis turned out in favour of the former. When Olweus summarised his results, the following picture emerged. The dramatis personae of the story are ‘potential bullies’ and ‘potential whipping boys’. Both may be either present or absent, thus yielding a total of four logical possibilities. Conceived in these terms, the explanatory task divides into two supplementary strands. The first task is to ascertain which personality traits are characteristic of the two types and to determine the factors that produce them. The second task is to account for how the presence of a ‘potential bully’, a ‘potential whipping boy’ – or their simultaneous presence – produces mobbing in a school class. This is what Dan Olweus attempted to do. He proposed an account where the central mechanisms of mobbing were taken over directly from aggression research – couched, for example, in terms of ‘aggression-inhibitory’ and ‘aggression-stimulating’ tendencies in the personality equipment of individuals. The first formulation of the concept of mobbing, and its subsequent evolution in public discourse, supported the notion that mobbing was a special case of aggression. In the previous section, we showed that this presented Dan Olweus

42 with opportunities as well as problems. It gave him the opportunity to transplant his research on aggression to the newly discovered phenomenon. At the same time, Konrad Lorenz’s view of aggression, upon which this conception of mobbing relied, was unacceptable to Olweus. In order to take over the term ‘mobbing’, his use of it must conform reasonably well to the way it was used in everyday speech, while at the same time avoiding the presuppositions inherent in that discourse. His subtle reformulation of the concept was a response to that im- passe. A similar problem recurred, however, in relation to the theoretical model. The Swedish discourse on mobbing between 1969 and 1973 had generated a series of entrenched lay explanations. If Olweus’ empirical inquiry was to be perceived as a contribution to the study of mobbing, rather than of aggression, those explanations could not be passed over in silence. They were consequently integrated, in the guise of ‘situational factors’, as hypotheses in the theoretical model. Despite the fact that Olweus’ results contradicted each and all of these explanations, his inclusion of them would, paradoxically, contribute to their perpetuation and diffusion. What we have referred to above as the counter-model brings together a set of commonly endorsed explanations. They were examples of common sense, and testing them against data can best be seen as an act of vigilance against spontaneous thinking. But we are dealing here with a specific version of spontaneous thinking, the contents of which were highly context dependent, for these particular lay ‘explanations’ took shape as a result of conceptual contagion and issue- symbiosis in Sweden of the late 1960s and early 1970s. Upon being incorporated in Olweus’ theoretical model, their link to a particular time and place is severed. When Dan Olweus began a new investigation of mobbing in 1983 – where a total of more than 150.000 Norwegian and Swedish pupils were investigated – the counter-model from the 1973 book reappeared.52 He investigates anew the familiar questions about the role of school size, class size, and physical deviance in producing mobbing (Olweus, 1986). Moreover, the new sample allowed Ol- weus to address another explanation produced by conceptual contagion and is-

43 sue-symbiosis in the early discourse on mobbing: the role of city size (Olweus, 1986:23). Just as in the 1973 book, Olweus’ results from the 1980s disprove each of the hypotheses pertaining to the counter-model. Needless to say, it is difficult to determine whether or not these old lay conceptions were still alive in Swedish collective conscience as of 1986. The author, at any rate, argues that they were. This is what he says, for example, about the idea that mobbing occurs because the perpetrators themselves are insecure:53

It is a common perception among psychologists and psychiatrists that individuals with aggressive and tough behaviour really and ‘deep down’ are anxious and insecure. The question as to the possible underlying insecurity in the perpetrators of mobbing has been meticulously tested in my own investigations with a variety of methods (including analysis of stress hor- mones and so-called projective tests). There is nothing in the results to support the current view. If anything, the analyses point in the opposite direction: the perpetrators either exhibit unusually low or approximately average degrees of anxiety and insecurity. (Olweus, 1986:34)

The argument is that there are lingering traces of the early discourse in the mid 1980s. Olweus makes similar remarks about the other entrenched explanations, and their inclusion in the counter-model comes out as an ongoing vigilance against spontaneous thinking. Note, however, that regardless of whether Olweus’ assessment of discursive remnants is accurate, his continuous mention of these ‘hypotheses’ passes them on as contenders to the title of best explanation. As we shall see next, this feature was further accentuated as Olweus’ works were transferred from a Swedish to an international context. Dan Olweus 1973 book on mobbing was translated into English in 1978, under the title Aggression in the Schools: Bullies and Whipping Boys. Some additions and revisions were made to the English edition – notably in chapters five, seven, eight, and nine – but by and large it is a translation rather than a new work (Olweus, 1978:xi). Before that, it had been published in Norwegian (1974) and

52 The massive scale of the investigation is explained by the fact that Olweus’ questionnaire, developed for the project, was used by the Norwegian state authorities – Kirke- og undervisningsdepar- tementet – in a nationwide campaign against mobbing (Olweus, 1986:11). 53 As described in section 3 above, playwright Ingrid Sjöstrand apparently introduced this notion to a larger audience in 1970.

44 in Danish (1975). Similarly, his book from 1986 was subsequently published in English, German, and French.54 It is possible that, to a limited extent, the Danish and Norwegian readership could have been aware of the lay explanations that evolved in the Swedish discourse between 1969 and 1973.55 But this was clearly not the case outside of Scandinavia. One consequence of this is that – to an Eng- lish, German or French reader – the boundaries between core and counter model become blurred. Hypotheses pertaining to the counter-model, and hence to the early Swedish discourse on mobbing, would appear to be as much part of Olwe- us’ theoretical model and research design as those that were derived from his research on aggression and belonged to the core model. The only difference between them is that so far, in Olweus’ own investigations, the former have turned out false. But since Olweus appears to regard them as important, other scholars are invited to test whether these, so far rejected, hypotheses might come out true in new inquiries and other populations. In a summary statement about state-of the-art research on childhood bullying, Gunilla Björk notes that, from an international perspective, Dan Olweus is the most prominent researcher in the field. ‘Several European projects also appear to be direct copies of his research’ (Björk, 1999:29). And to underscore her point, she goes on to quote another scholar’s survey of the field:

The common denominator, in the body of research we have referred to, is that it focuses on charting the prevalence of mobbing, and on investigating how it correlates with certain personality traits in victims and perpetrators and with certain organisational factors – such as school size and the com- position and group climate of the school class. (Fors, 1993:21).

If Björk and Fors are correct in their assessments, we have a situation where international scholars copy Olweus’ research and focus on the factors incorporated in his model, regardless of whether they belong to the core or the counter-model. This is why both ‘personality traits’ (core model) and ‘school size’ (counter- ––––––––– 54 Cf. Dan Olweus (1993) Bullying at school: what we know and what we can do; Dan Olweus (1994) Gewalt in der Schule: was Lehrer und Eltern wissen sollten - und tun können (see also later editions in 1995, 1996, 1997, 1999, 2002, 2004, 2006); Dan Olweus (1999) Violences entre élèves, har- cèlements et brutalités: les faits, les solutions.

45 model) are included as hypotheses in the majority of research projects on mobbing. Somewhat paradoxically, then, Olweus served as a vehicle of diffusion for the lay conceptions he set out to combat. Their prominence in the discourse meant that Olweus was obliged to take them into account. By doing so, however, he contributed to their perpetuation, de-contextualised them, and passed them on to scholars who were unaware of their origin in Swedish public debate between 1969 and 1973.

55 Peter-Paul Heinemann’s book on mobbing, which appeared in Swedish in 1972, was translated into Danish in 1972 and into Norwegian in 1973.

46 7. Conceptual entrepreneurialism and embedded discursive remnants: an analytical summary

In the previous chapter, we chartered how the concept of mobbing first entered the scientific literature. The task now is to pull these rather detailed analyses together in a more general argument about the interaction of lay and scientific discourse. The career of the concept of mobbing has some interesting properties, the significance of which go beyond the case at hand. When a social problem emerges on the public agenda, the community of social scientists are invited, nay obliged, to have a say on the matter. This can entail a wide range of activities, from interviews in the press to careful scientific scru- tiny of the phenomenon at hand. The case history of mobbing contains both elements. Journalists and playwrights asked psychologists, sociologists, and education scholars to give their opinion about the nature, causes, and consequences of mobbing. The latter had to relate to a concept that was not the carefully crafted product of research in their respective disciplines. Their analyses had to be based upon analogy or upon generalist knowledge incorporated in the textbooks of their respective disciplines. This, I argue, is what happened – in late 1969 and early 1970 – when playwright Ingrid Sjöstrand sought psychological advice in writing her play about mobbing. Sociologists Rolf Nielsén and Lars Stigendal did the same, only drawing upon sociological rather than psychological common knowledge (Nielsén & Stigendal 1973). The mechanism involved is quite general. Fa-

47 ced with a request to analyse an unknown entity, social scientists respond by of- fering applications of standard knowledge in their respective fields.56 The same type of mechanism operates, albeit in a more complex fashion, when scholars move into the field and address the phenomenon empirically. In such cases, textbook standards give too little guidance to be of any serious use. The researcher needs a transplantable model of a different kind: methods, theories, and concrete procedures employed in previous research, preferably of his or her own making. Dan Olweus illustrates this. His inquiries drew extensively on his own previous research on aggression. Yet, however undefined and poorly circumscribed the concept of mobbing may have been in general discourse, it was not devoid of content. Olweus would have to relate to that content if his research was to be credible as a contribution to the understanding of mobbing. His use of the concept must be sufficiently in line with current usage to prevent it from being aborted, and his research must, some way or the other, address entrenched lay explanations of the phenomenon. Similar restrictions, it may be noted, are imposed upon any scholar who projects her previous research unto a topic that leads a life of its own in public discourse. What interests us next is how such restrained dialogues between lay and scientific conceptions are played out. The fate of ‘mobbing’ can serve as exemplar in this respect as well. Although the details of this story are highly context dependent, and although the concrete development is governed by a case specific concatenation of mechanisms, I argue that it is possible to abstract and isolate mechanisms that are active across a range of cases. The way to get at them is to pay close attention to the case at hand. Consider first how Dan Olweus related to the lay concept of mobbing. The current conception of mobbing as a subtype of aggression gave Olweus the semantic foothold he needed to present his old methods and theories as answers to questions about mobbing. At the same time, the early image of mobbing carried

––––––––– 56 Note that such contributions blends with and contributes to the evolution and entrenchment of lay conceptions. The message of Ingrid Sjöstrand’s play – that mobbing occurs because, deep down, the perpetrators are insecure – no doubt bore the mark of psychological standard knowledge. Her play served to amplify this notion. If Dan Olweus’ assessment of Swedish collective consciousness was correct, this notion was still alive in the mid 1980s.

48 with it presuppositions that did not sit easily within his framework, e.g. a Loren- zian view of aggression and a root meaning emphasising group violence and all- against-one situations. Olweus wanted the label ‘mobbing’, but not everything that came with it. Just like a hermit crab coveting a new shell sometimes has to jostle its current inhabitant for possession, scholars who want to transfer their old studies to a new conceptual domain may face the task of suppressing established denotations and connotations. The history of social science harbours quite a few jousting crustaceans.57 We may describe Dan Olweus’ reformulation of the concept of mobbing as a case of conceptual entrepreneurialism. He took over an entrenched concept, while simultaneously transforming it so as to be compatible with research he had already conducted under a different label. Ultimately, his intervention changed the meaning of mobbing and made the concept incompatible with some of the core examples on which it first rested. Such a drastic manoeuvre, had it been proposed as a definition, would not have squared with current usage in 1973, and the reformulation may well have been aborted. Instead, the specification that mobbing was ‘repeated and durable’ negative actions was originally introduced as a temporary delimitation pertaining to a concrete study. Only later, by means of enveloping quotations, was this specification gradually turned into definitional criteria. Thus, in cases where a direct redefinition would go against the grain of established discourse, the safest route from collocation to formal equivalence goes via embedded quotations. Even when successful, however, conceptual entrepreneurialism comes with a price. A move into a new shell can release resources (e.g. research grants, attention in the press, involvement in government activities), but it involves the risk of becoming associated with its former resident. There may be lingering traces of old, suppressed denotations and connotations in public discourse. More importantly, the scholar who takes over an entrenched lay concept faces the challenge

––––––––– 57 Think, for example, of Durkheim’s usurpation of ‘anomie’ from its original meaning in Jean-Marie Guyau’s L’irreligion de l’avenir (1887). Or think of the very concept of sociology, and how various parties have endowed the term with a meaning that was congenial with what they were already doing under other labels.

49 of continually fending off whatever explanations and accounts have already accumulated in connection with it. This is what happened in the case of mobbing. In the Swedish discourse evolving between 1969 and 1973, a combination of conceptual contagion and institutionally backed issue-symbiosis generated and cemented an entire series of taken-for-granted explanations of mobbing: school size, class size, suburban and metropolitan life, and physical deviance in the victims (including ‘deviations’ such as being an immigrant or adopted child). Dan Olweus incorporated them all in his research design and theoretical model. Ta- ken jointly, they make up what we have referred to as his counter-model, as opposed to the core model derived from his research on aggression. Olweus’ integration of factors pertaining to core and counter-model in a single framework, and the diffusion of this model to other contexts (be it Sweden of the 1980s or the world of international scholars), had some curious effects. A common, even routine, heuristic gambit in social science is to replicate a previous study on a new population, sometimes adding a variable or two (Abbott 2004). This is what the international scholarly community did with Olweus’ work. They used his questionnaire and tested for the factors listed in his model, regardless of whether they belonged to core or counter-model. They thereby contributed to the dissemination and perpetuation of lay explanations that arose in a very specific time and place. Despite – or rather because of – Olweus’ continued vigilance against a historically situated version of spontaneous thinking, these discursive remnants were transformed into de-contextualised, abstract, and serious contenders. There is no reason to suppose any two conceptual careers to be identical. The- re is plenty of room for idiosyncrasies, and a good deal of context dependency. In addition, the case history of mobbing offers a rather slender diet of examples. On the other hand, this relatively simple and condensed case history has admitted a fairly detailed treatment, and some of the mechanisms invoked above are plausible candidates for generalisation. There are rewards in store for aspiring conceptual entrepreneurs, particularly if the concept to be claimed is prioritised on the public agenda. Staking that claim, however, is likely to involve concessions. In order to move into the new turf, the conceptual entrepreneur has to stay reasonably close to the entrenched mea-

50 ning of the concept.58 Depending on what that meaning is, some disciplines and some scholars are favoured over others. Yet even when there is a good fit, where a scholar’s usual research practices can be dressed up in the new label with relative ease, previous discourse is likely to contain elements that don’t quite square. The conceptual entrepreneur thus has to relate to a partially alien body of examples, accounts, explanations, and implicit or explicit definitions. To the extent that this involves slicing the phenomenon differently from established discourse, suc- cess hinges upon a smooth introduction and acceptance in subsequent enveloping quotations. Even with such conceptual modifications, there is still an obligation to take on the existing body of explanations – they are, after all, the best contenders so far – if only to prove them wrong. This move, however, easily teams up with another set of mechanisms (de-contextualisation, normal social science heuristics, enveloping quotations) to perpetuate and disseminate the old stock of explanations.

––––––––– 58 ‘Degree of entrenchment’ obviously suggests itself as a key variable in studies of conceptual careers.

51 8. Coda: a note on ’mobbing’ since the mid 1980s

For the purpose of theorising conceptual evolution in the social sciences, the analytical summary above is a suitable final chord. I have two reasons for not making it the endpoint of the inquiry. First, my preoccupation with the early history of mobbing, and with aspects of Dan Olweus’ work, may invite illicit extrapola- tions. The period between 1969 and 1973 was chosen precisely because it was sufficiently simple and condensed to allow us to study some basic mechanisms involved in conceptual evolution. This same feature, however, makes it an ina- dequate portrait of how the concept has fared in the recent past, notably since the mid 1980s. This is a more complex story. Influential as Dan Olweus has been in shaping lay, administrative, and scientific discourse, his legacy falls far short of hegemony. For the sake of historical accuracy, then, a few words must be said about subsequent developments. Second, an outline of these developments, however sketchy, can serve as a laboratory where we can try out some of our analytical propositions in a more complex environment – and suggest a direction for where to look for a few more. We commence our inquiry on familiar territory: the publication in 1986 of the Swedish edition of Dan Olweus’ Bullying at School. At this point, mobbing was defined as a situation where ‘one or more individuals are subjected to negative actions, on several occasions and over an extended period of time, by one or more individuals’. We have already spoken our piece about how the definition ended up there. What interests us now is what it made possible. Until the mid 1980s, the discourse on mobbing had focused on children and on school settings. There are exceptions. Early on, and no doubt in line with Heinemann’s Lorenzian image of mobbing, there were examples of groups of

52 pupils mobbing their teacher. When – in 1973 – female employees in Stock- holm’s municipal administration were taunted by colleagues for their ‘bourgeois custom’ of wearing skirts, this was described as ‘political mobbing’.59 And in 1977, female theology students, harassed by their male fellow students, described their predicament in terms of mobbing.60 Extensions of this type are no doubt to be expected when a concept has been around for a while in everyday speech. Current semantics did not strictly disallow them. To the extent that mobbing was modelled upon the Lorenzian image of herbivore birds attacking a predator, the concept seemed more readily applicable to child than to adult behaviour. On the other hand, the term could also refer to the isolation, ostracism, and ‘freezing out’ of a peer. And if mobbing could be something other than physical violence, then it could also be transferred to adult behaviour. But Peter-Paul Heinemann’s examples were taken from the world of children, the press emphasised the situation in the schools, and the organisations and institutions that picked up and promoted the category were child or school related. Moreover, the research presented in Dan Olweus’ 1973 book on mobbing was exclusively concerned with schoolboys. Applications to adults remained exceptions, but they became increasingly common in the 1980s, when ‘mobbing’ once again became a vogue word. There is no doubt that ‘schoolchildren’ was the implicit reference class for Dan Olweus’ definition of mobbing. The definition itself, however, is stated in very abstract terms. As long as there are repeated and durable negative actions directed against one or more individuals, the victims and perpetrators need not be children and the setting need not be schools. This feature was an invitation to new conceptual entrepreneurs, foremost among them psychologist Heinz Ley- mann. Prior to the mid 1980s, Heinz Leymann had published in a wide variety of areas: the psychology of victimization, stress research, the psychology of work place organisation and work environment. These writings converged in a general interest for the psychology of working life, and Leymann was connected both

––––––––– 59 Svenska Dagbladet 16 June 1973. 60 Svenska Dagbladet 2 April 1977

53 with the Swedish Work Environment Authority and with the National Institute for Working Life. In the years to come, he was to emerge as the leading expert on mobbing in the workplace. He and others firmly established mobbing as a scientific category embracing children and adults alike. There are similarities between Leymann’s and Olweus’ respective paths to the concept. Leymann, too, had to relate to entrenched discourse, and he too was hesitant about the appropriateness of using the word mobbing. If Olweus had ‘aggression’ as his preferred concept, Leymann had ‘psychological violence’. This was the term used in a research report for the Swedish Work Environment Authority that he co-authored in 1985 (Gustavsson & Leymann, 1985). Most Swedish newspapers wrote about the report, and it was taken up in national television in the Scandinavian countries. Union and company health service representatives contacted the authors and offered to participate in their research (Le- ymann, 1986: 11f). When the first results from that research were published two years later, Leymann was wavering about the terminology:

What I intend to call by its true name – psychological violence in working life – is popularly called ‘mobbing’. And this word refers to just about anything that can go on in human interaction. There is a danger lurking here: if we don’t describe what actually happens when someone ‘mobs’, if we don’t learn to describe the event in concrete terms, we may end up in the unfortunate situation where anyone can accuse anyone of mobbing. A vogue word that is elevated to the status of diagnosis can hit harder than intended. Mobbing is a vague word, which says little more about what has happened than that the person who uses it feels hurt and angry. Such a word is really just an invective, a denunciation. If it gets the status of a diagnosis, there is a risk of grave injustices. (Leymann, 1986:14)

Leymann was obviously uncomfortable with using the term mobbing. He goes to some length to show that it harbours vagueness and vagaries, and he makes it no secret that he regards ‘psychological violence’ to be the more accurate concept. But the title of the booklet nevertheless translates as Mobbing among adults: On psychological violence in working life (Leymann, 1986). The preferred concept was thus relegated to the subtitle, and it gradually disappeared from the titles of subsequent publications (cf. Leymann 1991a; 1991b; 1991c;1992a; 1992b). But this is not to say that the notion of psychological violence was abandoned. On

54 the contrary: just as mobbing was aggression to Dan Olweus, mobbing was psychological violence to Heinz Leymann. ‘Mobbing’, as Leymann used the term, became a new label for psychological violence. This is conceptual entrepreneurialism of the same kind as we have analysed above. So why did Heinz Leymann trade psychological violence for mobbing? The answer can be read from between the lines in the quotation above. Once again, ‘mobbing’ was in vogue. This was how people talked about the phenomena he wished to describe. Ingela Thylefors, another psychologist who began to study adult mobbing in the 1980s, described the situation as follows:

In the last few years, there has been a discussion in mass media and in the trade press about individuals who, in their working life, are frozen out by their peers. The debate escalated with the formation of an association called Exodus, in December 1983. The initiative to form this interest organisation came from people who had been subjected to various forms of harassment and ostracism in their respective work places. Several synonymous concepts are used in the debate, e.g. ostracism, harassment, mobbing, marginalisation, and psychological violence. The concepts are defined somewhat differently. But the core element in most of them is nevertheless that they refer to cases where one or more individuals are subjected to negative actions, several times and for an extended period of time, from one or more individuals. (Thylefors, 1987: 20)

Three aspects are worth emphasising at his juncture. First, the concept of mobbing had undergone a new cycle of entrenchment. It was gradually transferred to adults and conditions in work organisations in general discourse – much due to institutional backing, in the guise of newly formed organisations. Second, Dan Olweus’ definition of mobbing was instrumental in that transfer. It was abstract enough to allow the concept of mobbing to be applied to adults. One of Heinz Leymann’s measures to prevent mobbing from becoming an empty accusation was to introduce the specification that the psychological violence must occur ‘regularly and for an extended period of time’ (Leymann, 1986:14). Indeed, if Thylefors’ description is correct, Olweus’ criteria for mobbing had entered the definitions of a series of adjacent concepts. Third, mobbing had become a more heterogeneous category. The extension of the term to adults allowed comparisons between two forms of mobbing, descri-

55 bed in two different literatures. Dan Olweus’ research focused on children and schools, Heinz Leymann on adults and work life. Fifteen years later, sociologist Björn Eriksson was to exploit this comparative horizon as a levy to ask entirely different questions about mobbing. Granted that mobbing occurs in schools and workplaces, he asked, where else do we find it? Not everywhere, apparently. But if mobbing has its habitat in particular arenas, Eriksson argued, then Dan Olweus’ account, in terms of stable personality traits in victims and perpetrators, can at best be a partial explanation. We carry our personality traits with us all the time, so if mobbing only occurs in particular loci, something more is needed to explain it (Eriksson, 2001). Let us return, however, to the mid 1980s. So far, we have had no difficulties in spotting some of the same mechanisms that we analysed above. To this list, we can add the impact of translations and enveloping quotations. Heinz Ley- mann soon became available in German. His works have been quoted frequently both in Sweden and abroad, which has contributed to the diffusion of his notion of mobbing as psychological violence in the workplace.61 But as we shall see, things start to get a bit messy at this point. There is little point in shoehorning the sequence of events into a preconceived set of mechanisms. We do best, rather, to indicate where it would be profitable to look for new ones. Both Dan Olweus’ and Heinz Leymann’s books on mobbing appeared in Swedish 1986. Both stirred considerable interest. They triggered separate but interrelated developments in science, lay discourse, law, and administration. Each unfolded over time according to its own logic; all interacted with all. This is too tangled a story to be un-

––––––––– 61 ‘Psychoterror am Arbeitsplatz’, which occurs in the subtitle of the German edition of Leymann’s work, is the German rendering of ‘psychological violence in the work place’. The phrase recurs in a long array of works – cf. Thomas, 1993; Huber, 1994; Sellier, 1997; Arentewicz, 2003; Hermans, 2004; Kolodej, 2005. This testifies that the mechanism of enveloping quotations is in operation. The German reception of Leymann paved the way for Dan Olweus. But the co-existence of two rather different conceptions of mobbing in Germany has been a source of confusion. As we would expect, this has led some scholars to search for a pure source in Leymann’s and Olweus’ references – and have found Konrad Lorenz at the end of the line. Roth (2002), for example, is a somewhat bi- zarre attempt to pursue the concept and theory of mobbing beyond Lorenz, to the ethologist Rudolf Bilz. Upon learning that Lorenz was not the first ethologist to use the concept of mobbing, Roth draws the amazing conclusion that Bilz’s writings is the true source of ‘mobbing’ in the modern sense, and that we would profit from applying his ideas on mobbing animals directly to the study of mobbing in the workplace. This should be a sobering reminder, especially to sociologists, that an unhistorical search for origins can produce nonsense.

56 ravelled here. But it is a good place to start looking if we want to expand the re- pertoire of mechanisms bearing upon travelling concepts. A few notes on the legal developments can serve as illustration. Leymann’s writings linked mobbing to state regulations of the work environment. Mobbing was suddenly relevant to the Occupational Safety and Health Act. Mobbing among school children also gained legal relevance. During the 1980s and 1990s, we find several cases where perpetrators of mobbing stand tri- al. Mobbing was not a legal category in itself: the defenders were charged with assault, sexual harassment etc. The discourse on mobbing had nevertheless brought these behaviours to the fore, and had, in that sense, contributed to a juridifi- cation of the school environment. In the years around the millennium, however, there was a momentous shift from criminal to civil law. A former schoolgirl and her mother took a Swedish municipality to court, arguing that it was guilty of criminal negligence in allowing the girl to be mobbed for several years, and cla- iming indemnity for a sum total of 657102 Swedish crowns. Instead of persecu- ting individual perpetrators under criminal law, mobbing was made an issue of tort law, invoking also sections of the School Act and of the Occupational Safety and Health Act (UFB6, 2005/06:628ff). After winning a partial victory in the di- strict court, the plaintiffs lost in appeal court, and that decision was confirmed – though not unanimously – in the Supreme Court. Even so, mobbing was established as a legally relevant category. In the verdict, dating from 2001, the Swedish Supreme Court noted:

In the last decades, and especially during the last fifteen years, mobbing problems have become visible to an unprecedented extent. It is, not least, the serious problem which mobbing poses in the schools that has been brought to the fore. This is reflected in the laws regulating the schools, which emphasise the schools’ far-reaching responsibility to intervene against all forms of harassment or injurious treatment. (UFB6 2005/06: 639)

The legal system thus had to handle issues of definition, of terminology, of causal attribution, of culpability and indemnification, and it had to do so from the point of view of the existing legal framework and praxis. These are strictly legal matters, and they are worked out according to the logic of law. But there is obvi-

57 ous interaction with scientific discourse and administrative practice. First of all, mobbing became a legal issue much as a consequence of the scientific attention it received in the mid 1980s.62 Second, Swedish scholars who has written about mobbing have tended to absorb legal and administrative language, e.g. ‘injurious special treatment’. Third, it is sometimes necessary to relate to current research in legal assessments. In the case referred to above, for example, the key issue – according to the Supreme Court – was whether her school had done everything that could reasonably be expected of it to prevent the girl from being mobbed. Court proceedings can thereby become the battleground for competing theories. In the case at hand, the plaintiffs – supported by a representative from the organisation Föreningen mot mobbning (the Association Against Mobbing) – had argued that the school had dealt with the situation in a way that was in breech with scientific theories. The Supreme Court, however, dismissed this argument:

It cannot be an unconditional requirement that the measures taken by the school must have been successful. When we are dealing with a complica- ted social and psychological phenomenon, as in the case of mobbing, it cannot be required, in an assessment of a claim for damages, that the school must have applied a particular theory or a particular action programme and that the measures must have been introduced in a certain order. (UFB6 2005/06:640)

A much more detailed study would be needed, were we to give an accurate portrait of these and subsequent legal developments. As far as it goes, however, the account above indicates that there is interaction between science and law. We should be looking at how this interaction unfolds over time. The legal framework, moreover, significantly shapes local administrative procedures. As ‘mobbing’ gained in legal relevance, it became an increasingly important administrative category. Swedish schools are required to have action ––––––––– 62 The Swedish School Law of 1985 was repealed and amended in 1990, and the proposition presented to the Swedish Riksdag makes it clear that ‘all forms of mobbing must be countervailed’. Cf. Proposition 1990/91:18, p.30. It is worthwhile to note, moreover, a somewhat more indirect effect of the body of scientific literature. If a plausible impression can be created that we know the mechanisms of mobbing and can prevent it, an omission to do so comes out as an action. In Hermann Lübbe’s words, this amounts to a Zurechnungsexpansion, an expansion of the realm of

58 plans against injurious treatment. This is how Dan Olweus’ definition of mobbing moved into local school documents. The schools, in turn, are supervised. Between 20 August 2005 and 20 August 2006, 193 named schools from 116 municipalities were reported to the Swedish Board of Education for not taking proper measures to prevent mobbing.63 The pressures thus imposed on the schools create a market for consultants, education, expertise, reports, and scientific investigations.64 There is no doubt that there is interaction between science and

causal attribution (Lübbe 1994a; Lübbe 1994b. Only upon such an expansion – often powered by scientific inquiry – can omission be seen as negligence, and only then is tort law applicable. 63 Aftonbladet 21 August 2006. 64 As an example, we may cite Gunilla Björk’s account of how her own research project began. In 1983/84, she says, there a large survey on mobbing was conducted in Swedish and Norwegian schools. Dan Olweus headed the survey project, and the first results ended up in his 1986 book. Gothenburg was one of three Swedish municipalities to participate. When the results began coming in, the Gothenburg School Board decided to install a resource group as a supplement to the ordinary work against mobbing. Björk became part of that group:

It was clear already after six months that the schools took great interest in the resource group. The services that were most in demand were consultancy and education. In all this educational work we made frequent use of Olweus’ large survey. The ‘mobbing project’ had after all emerged in response to it, and the demand for our services was enormous. (Björk, 1999, 16)

A soaring demand for education, in conjunction with the fact that Olweus’ inquiry was the only available large-scale investigation, pushed Björk and her colleagues to envelope Olweus’ views when teaching and consulting. Later on, when Björk had her own research to go on, the educational contents could begin to slide. When demand for mobbing consultants and education increases, there is room for more actors in the market. Note that an investigation of the interplay between legal pressures, administration, and science should not be limited to Sweden. To underline that this is an international phenomenon, it is useful to quote the order form for the “Revised Olweus Bully/Victim Questionnaire, PC Program and Teacher Handbook” (2001), so far available in English, Swedish, and Norwegian:

The costs for use of the Questionnaire for practical registration and/or intervention purposes (maybe with involvement on the part of local or central school authorities, parent/teacher associations etc) are as follows: * for use in 1-2 schools, US $ 140 (or £90) * for use in 3-5 schools, US $ 250 (or £160) * for use in 6-10 schools, US $ 375 (or £240) * for use in 11-20 schools, US $ 625 (or £400) * for use in 21-30 schools, US $ 890 (or £560) * for use in 31-40 schools, US $1125 (or £710) * for use in 41-50 schools, US $ 1400 (or £890) * for use in more than 50 schools, a special agreement (with discount) has to be made with me. Also in these cases, I would very much appreciate a brief description of the project or context in which the Questionnaire is to be used (e.g., on the back of the enclosed Order Form or in a separate letter).

59 administration: the question is how it developed over time, and what consequences it has had. Last but not least, we should be looking at how mobbing has fared in lay discourse since the mid 1980s. When the term mobbing first appeared in the late 1960s and early 1970s, it struck a raw nerve. The examples that came with it organised experience in a new way: they were uncannily familiar. Mobbing, it would seem, had been with us from time immemorial, but only with the aid of the new concept could they be made visible. Gradually, through Dan Olweus’ subtle reformulations, the concept drifted away from its ethological background. He uprooted taken-for-granted and institutionally embedded knowledge about the phenomenon. This may have cooled off some of the first political enthusiasm, but the concept was there to stay. It purported to name a phenomenon, the consequences of which were so dire that it would seem insulting to abolish the term. A host of organised activities had also been built up around the category of mobbing. One important issue, then, is how the entrenchment of the term – and its scientific, legal, and administrative status – has affected lay conceptions of mobbing. Conversely, we may ask how the evolution of lay conceptions works to change or split the concept of mobbing. To borrow a key phrase from Ian Hack- ing: how do names interact with those who are named? (Hacking, 2006:3) It is useful to stay with Hacking for a while. He has set out to design a framework for the analysis of what he refers to as ‘the looping effect’ and ‘making up people’.65 Whether the purpose was to control, to organise, to help, or to emulate, science has turned each of them into an object of inquiry and treated them as kinds of people:

Previous research, Olweus informs us, has shown a fifty percent reduction of mobbing problems in school where the questionnaire and its associated intervention techniques have been used. That is no doubt a strong motive for buying it. The order form also gives a sense of the scale of the operation. There are separate packages that can be bought by researchers, on the condition that they state the aims of their research and report the results back to Dan Olweus. Thus, pressure on the schools to take (legally or morally) proper courses of action generates a demand for education and inquiries; results from these inquiries feed back into the research apparatus. Established actors have an advantage, but there is room in the market for newcomers. 65 Cf. also Hacking (1995a; 2004). For detailed empirical illustrations, see Hacking (1995b; 1999).

60 We think of these kinds of people as given, as definite classes defined by definite properties. As we get to know more about these properties, we will be able to control, to help, to change, or to emulate them better. But it is not quite like that. They are moving targets because our investigations interact with the targets themselves, and change them. And since they are changed, they are not quite the same kind of people as before. The target has moved. That is the looping effect. Sometimes our sciences create kinds of people that in a certain sense did not exist before. That is making up people. (Hacking, 2006: 2)

Does this apply to mobbing? Preliminary evidence suggests that it does. Mob- bing was a lay category – or, better, a lay translation of an ethological concept – before it became an object of scientific inquiry. As kinds of people, the victims of mobbing were endowed with a set of properties right from the start. The kind was entrenched, but the properties were altered, when victims and perpetrators became objects of psychological inquiry. In this sense, scientific intervention really did ‘make up people’. New slots were created in which to fit and enumerate people. Perpetrators and victims were counted and quantified, norms were imposed to regulate them, correlations with a host of variables were tried for significance, and they were dealt with as categories in bureaucratic practice. Yet we still need a clearer notion of what it means when science is said to make up people. Consider the following alternative formulations, paraphrased from Hack- ing:

(A) There were no victims of mobbing in 1968; there were many in 1998. (B) In 1968 this was no a way to be a person, people did not experience themselves in this way, they did not interact with their friends, their families, their employers, their counsellors, in this way; but in 1998 this was a way to be a person, to experience oneself, to live in society. (C) The mobbing victim, as a kind of person, did not exist in 1968, it did exist in 1998.

Formulation A is patently false. There is overwhelming evidence that, whichever way you cut it, mobbing behaviour has been with us long before that. The crux of formulation C is that one of its interpretations is A. If we steer away from that explication, however, we end up close to B. Opting for formulation B, then, we must be open to the possibility that the victim of mobbing is a moving target.

61 In particular, an empirical investigation of how the concept of mobbing evolved from the 1980s onwards should pay attention to the cultural image of mobbing developing in associations formed by mobbing victims. We have already mentioned the organisation Exodus, formed in 1983 by people who had been harassed and ostracised in their workplaces. Föreningen mot mobbning (the Asso- ciation Against Mobbing) was formed in 1992, and a similar organisation – Fri- ends – in 1997. If being a victim of mobbing was sufficient grounds for organisation, it was also a candidate master status (Hughes 1993). To the extent that others think of this ‘kind of people’ as a given, as a definite class defined by it properties, we may expect these organisations to launch attempts to reclaim their identity.66 Such attempt, in turn, must interact with scientific, administrative, and legal developments. The task is to see how that interaction unfolds over time.

––––––––– 66 This may be a particularly pressing issue in the case of mobbing. To bring this out, let us compare mobbing to the cultural image of stress, as described by Andrew Abbott. The concept of stress, if not the word, goes back to the beginning of the 20th century. It is the image of an individual, ‘at once damaged by society and maladjusted to it’ (Abbott, 2001:42). This notion introduces a systematic duality in the cultural image of stress. And, Abbott goes on to say, ‘ this image had (and has) its great meaning for us precisely because it binds anxiety and mind cure with social adjust- ment and performance’:

It might seem that there is damage from the individual’s point of view and maladjustment from the society’s. But the individual is always worrying that ‘maybe it’s me who is to blame,’ while few if any societies completely ignore the burdens they place on individuals. (Abbott, 2001:43)

Thus, the constitutive ambiguity of the concept allows causal attribution to go either way. Interpreta- tions of particular cases are liable to be contested terrain, especially when parties with opposing interests are to collaborate in producing an account. There is a similar constitutive ambiguity to ‘mobbing’. In a case where an individual is ostracised by his peers, we may attribute this to aggressive tendencies in the peer group or to personality traits in the victim. As in the case of stress, this duality – and the opposing causal attributions it allows – is a feature that adds to the meaning of the concept. It is also, however, a threat of stig- matisation to those who are the victims of mobbing. Associations organising mobbing victims therefore have a strong incentive to promote a concept of mobbing that hands over the privilege of definition to the victim. This is exactly what the legal apparatus must avoid. Conceptual evolution is drawn into this force field as well.

62 References

Abbott, Andrew (2001) The Duality of Stress, In: Chaos of disciplines, Chicago: Chicago University Press

Abbott, Andrew (2004) Methods of discovery: Heuristics for the social sciences, New York: Norton

Angelöw, Bosse & Jonsson, Thom (1990) Introduktion till socialpsykologi, Lund: Studentlitteratur

Arentewicz, Gerd (2003) Arbeitsplatzkonflikte. Mobbing als Psychoterror am Arbeitsplatz. Ursachen, Folgen und Formen der Hilfe, Frankfurt a/M: Lang

Björk, Gunilla (1999) Mobbning – en fråga om makt?, Lund: Studentlitteratur

Eriksson, Björn (2001) Mobbning: en sociologisk diskussion, Sociologisk Forsk- ning, vol.38 (2): 8-43.

Eriksson, B., Lindberg, O, Flygare, E & Daneback, K. (2002) Skolan - en arena för mobbning, Stockholm: Liber

Fors, Zelma (1993) Obalans i makt: fallstudier av barnmobbning, Göteborg: Va- sastadens bokbinderi

Forsman, Arne (2003) Skolans texter mot mobbning – reella styrdokument eller hyllvärmare?, Luleå: Institutionen för lärarutbildning, Luleå tekniska universitet

Frank, Denis (2005) Staten, företagen och arbetskraftsinvandringen: en studie av invandringspolitiken i Sverige och rekryteringen av utländska arbetare 1960- 1972, Växjö: Växjö University Press

63 Gustavsson, Bo-Göran & Leymann, Heinz (1985) Psykiskt våld i arbetslivet: två explorativa intervjuundersökningar, Stockholm: Arbetarskyddsstyrelsen

Hacking, Ian (1995a) The looping effects of human kinds. In D. Sperber, D Pre- mack & A. Premack (eds.) Causal Cognition: An Interdisciplinary Approach, Oxford: Oxford University Press.

Hacking, Ian (1995b) Rewriting the soul: multiple personality and the sciences of memory, Princeton: Princeton University Press

Hacking, Ian (1999) Mad travellers: reflections on the reality of transient mental illnesses, London: Free Association Books

Hacking, Ian (2004) Making up people. In: Historical Ontology, Cambridge, MA: Harvard University Press

Hacking, Ian (2006) Kinds of people: Moving targets, The Tenth British Acade- my Lecture, http://www.britac.ac.uk/pubs/src/_pdf/hacking.pdf

Heinemann, Peter-Paul (1969) ’Apartheid’, Liberal debatt, 22 (2) 3-14.

Heinemann, Peter-Paul (1972) Mobbning: gruppvåld bland barn och vuxna, Stockholm: Natur och kultur

Hermans, Arndt & Krings, Elmar (2004) Praktische Mobbing-Prävention. Stopp dem Psychoterror am Arbeitsplatz!, Norderstedt: Books on Demand GmbH

Holmberg, Carl (1998) Längtan till landet: Civilisationskritik och framtidsvisio- ner i 1970-talets regionalpolitiska debatt, Göteborg: Historiska institutionen, Univ.

Huber, Brigitte (1994) Mobbing: Psychoterror am Arbeitsplatz, Niedernhausen/Ts.: Falken

Hughes, Everett C. (1993) Dilemmas and contradictions of status. In: The Socio- logical Eye: Selected Essays, New Brunswick: Transaction

Kolodej, Christa (2005) Mobbing: Psychoterror am Arbeitsplatz und seine Bewältigung. Mit zahlreichen Fallbeispielen, Wien: WUV

Leymann, Heinz (1986) Vuxenmobbning: om psykiskt våld i arbetslivet, Lund: Studentlitteratur

Leymann, Heinz (1991a) Vuxenmobbning på svenska arbetsplatser: en rikstäck- ande undersökning med 2438 intervjuer: delrapport 1, Solna: Arbetarskyddssty- relsen

Leymann, Heinz (1991b) Manligt och kvinnligt vid vuxenmobbning: en rikstäck- ande undersökning med 2438 intervjuer: delrapport 2, Solna: Arbetarskyddssty- relsen

Leymann, Heinz (1991c) Psykiatriska problem vid vuxenmobbning: en rikstäck- ande undersökning med 2438 intervjuer: delrapport 3, Solna: Arbetarskyddssty- relsen

Leymann, Heinz (1992a) Lönebidrag och mobbad: en svag grupps psykosociala arbetsvillkor i Sverige, Solna: Arbetarskyddsstyrelsen

Leymann, Heinz (1992b) Mobbning av föreståndare inom två barnomsorgsom- råden, Solna: Arbetarskyddsstyrelsen

Leymann, Heinz (1993) Mobbing: Psychoterror am Arbeitsplatz und wie man sich dagegen wehren kann, Reinbek bei Hamburg: Rowohlt-Taschenbuch- Verlag

65 Leymann, Heinz ed. (1995) Der neue Mobbing-Bericht: Erfahrungen und Initiativen, Auswege und Hilfsangebote, Reinbek bei Hamburg: Rowohlt

Leymann, Heinz & Zapf, Dieter (1996) Mobbing and victimization at work, Letchworth, Herts.: Psychology Press

Leymann, Heinz (1996) Mobbing, la persécution au travail, Paris: Seuil

Lorenz, Konrad (1963) Das sogennante Böse: Zur Naturgeschichte der Aggression, Wien: Borotha-Schoeler Verlag

Lorenz, Konrad (1968) Aggression – dess bakgrund och natur, Stockholm: Nor- stedt

Lorenz, Konrad (2002) On Aggression, London: Routledge

Lübbe, Hermann (1992a) Erfahrungsverluste. Lebensvorzuge und Lebenswelt- ferne der Chemie, In: Jürgen Mittelstraß & Günther Stock (ed.) Chemie und Geisteswissenschaften: Versuch einer Annäherung, Berlin: Akademie Verlag

Lübbe, Weyma (1992b) Handeln und Verursachen: Grenzen der Zurechnung- sexpansion, In: Weyma Lübbe (ed.) Kausalität und Zurechnung: Über Verant- wortung in komplexen kulturellen Prozessen, Berlin: Walter de Gruyter

Lundgren, Ulf P. (1999) Ramfaktorteori och praktisk planering, Pedagogisk forskning i Sverige, vol.4 (1): 31-41.

Moscovici, Serge (1961) La psychanalyse, son image et son public: Étude sur la représentation sociale de la psychanalyse, Paris: Presses Universitaires de Fran- ce.

Nilsén, Rolf & Stigendal, Lars (1973) Varför mobbning? Gruppvåld ur ett sam- hällsperspektiv, Vällingby: M & B fackboksförlaget

Nilsson, Åke (2004) Efterkrigstidens invandring och utvandring, Stockholm: Statistiska centralbyrån

Olweus, Dan (1969) Prediction of Aggression: On the Basis of a Projective Test, Stockholm: Scandinavian Test Corporation

Olweus, Dan (1970) Aggression from the perspective of personality psychology, Reports from the Institute of Psychology, No. 2, University of Bergen, Norway

Olweus, Dan (1973) Hackkycklingar och översittare: Forskning om skolmobbning, Stockholm: Almqvist & Wiksell

Olweus, Dan (1974) Hakkekyllinger og skolebøller: forskning om skolemobbing, Oslo: Cappelen

Olweus, Dan, (1975) Hakkekyllinger og skolebøller: forskning om skolemobning, København: Gyldendal

Olweus, Dan (1977) A critical analysis of the ’modern’ interactionist position, In: David S. Magnusson (ed.) Personality at the Crossraods: Current issues in Interactional Psychology, Hillsdale, N.J.: Erlbaum

Olweus, Dan (1978) Aggression in the schools: Bullies and Whipping Boys, Washington : Hemisphere publ.

Olweus, Dan (1986) Mobbning: vad vi vet och vad vi kan göra, Stockholm: Li- ber

Olweus, Dan (1994) Gewalt in der Schule: was Lehrer und Eltern wissen sollten - und tun können, Bern: Huber

67 Olweus, Dan (1999) Violences entre élèves, harcèlements et brutalités: les faits, les solutions, Paris: ESF éd

Panelius, Olav & Steinby, Torsten (1970) Svenska modeord, Lund: Gleerup

Perrin, Porter G. (1955) ‘Pecking Order’ 1927-54, American Speech, vol.30 (4): 265-268

Pikas, Anatol (1975) Så stoppar vi mobbning! Rapport från en antimobbnings- grupps arbete, Stockholm: Prisma

Pikas, Anatol (1998) Gemensamt-bekymmer-metoden: Handbok för ett para- digmskifte i behandling av skolmobbning, Uppsala: AMA dataservice

Roland, Erling (1983) Mobbning: Att förebygga och bearbeta gruppvåld bland barn och ungdom, Malmö: Liber

Roth, Wolfgang (2002) Mobbing – Herkunft des Begriffs und sein theoretischer Hintergrund, Gruppendynamik und Organisationsberatung, vol.33 (2): 197-211

Schjelderup-Ebbe, Thorleif (1913) Hönsenes stemme: Bidrag til hönsenes psykologi. Naturen: populaervitenskapelig tidsskrift 37: 262-276

Schjelderup-Ebbe, Thorleif (1921) Beiträge zur Biologie und Sozial- und Individualpsychologie bei Gallus domesticus, Greifswald: Adler

Schjelderup-Ebbe, Thorleif (1922) Beiträge zur Sozialpsychologie des Haushuhns. Zeitschrift für Psychologie und Physiologie der Sinnesorgane, vol.88: 225-252.

Schjelderup-Ebbe, Thorleif (1924) Zur Sozialpsychologie der Vögel. Zeitschrift für Psychologie und Physiologie der Sinnesorgane, vol.95: 36-84

68 Sellier, Florentine (1997) Mobbing – Psychoterror am Arbeitsplatz – Phänomen, Folgen und Analyse des Traumaverlaufs, Diplomica

Tanaka, Takeo (2001) The identity formation of the victim of ‘shunning’, School Psychology International, vol.22 (4): 463-476.

Thylefors, Ingela (1987) Syndabockar: om utstötning och mobbning i arbetslivet, Stockholm: Natur och kultur

Thomas, Rudolf (1993) Chefsache Mobbing: souverän gegen Psychoterror am Arbeitsplatz, Wiesbaden: Gabler

Official parliamentary publications and other official documents Skolöverstyrelsen (1973) Mobbning i skolan: Handledarhäfte, Stockholm: SÖ

Skolöverstyrelsen (1973) Mobbning i skolan: Elevhäfte. Stockholm: SÖ

SOU 1974:53 Skolans arbetsmiljö: Betänkande avgivet av Utredningen om skolans inre arbete – SIA, Stockholm: Allmänna förlaget

SOU 2003:49 Adoption - till vilket pris?, Stockholm: Fritzes offentliga publika- tioner

UFB6 2005/06, Rättspraxis i skolan, Norstedt juridik

Student essays Abrahamsson, M., Cederström, N., Sjöberg, B., Sjöström, S. & Tujus, M. (1973) Mobbning som företeelse och som föremål för studiecirkelverksamhet. Pedago- giska institutionen vid Stockholms universitet

69 Andersson, C. (1974) Behandling av tre fall av mobbning i skolan. Rapport i samband med antimobbningsgruppens arbete. Institutionen för pedagogik vid Uppsala universitet

Brandt, S, Broberg, A. & Strandell, M (1972) Ett försök till analys och behandling av kamratrelationerna i några klasser på grundskolans mellanstadium. Pe- dagogiska institutionen vid Göteborgs universitet

Christensson, G. (1974) Behandling av två fall av mobbning i skolan. Rapport i samband med antimobbningsgruppens arbete. Institutionen för pedagogik vid Uppsala universitet

Granath, T, Hedblom, M, Hom, M. & Sundén B.-M. (1973) Mobbning i skolan, Institutionen för pedagogik vid Uppsala universitet.

Gällingsjö, A.& Hallmo, M. (1971) Mobbning – ett aktuellt skolproblem, Social- högskolan i Örebro.

Hansson, B. & Ödman, E. (1973) Mobbning – Ett försök till analys och upptäckt i några klasser på grundskolans mellanstadium, Pedagogiska institutionen vid Göteborgs universitet

Hillner, S. (1974) Åtgärder mot mobbning: Analys av sju fall, Institutionen för pedagogik vid Uppsala universitet

Holmberg, L. (1974) Behandling av två fall av mobbning i skolan. Rapport i samband med antimobbningsgruppens arbete. Institutionen för pedagogik vid Uppsala universitet

Jeppson, M., Palm, E. & Unander, C. (1973) Mobbning mot lärare: en grupp lä- rares syn på tre hypotetiska fall, Institutionen för tillämpad psykologi vid Lunds universitet

70 Larsson, K. (1974) Behandling av tre fall av mobbning i skolan. Rapport i samband med antimobbningsgruppens arbete. Institutionen för pedagogik vid Upp- sala universitet

Nilsson, A. (1974) Behandling av två fall av mobbning i skolan. Rapport i samband med antimobbningsgruppens arbete. Institutionen för pedagogik vid Upp- sala universitet

Norrman, A. (1974) Behandling av två fall av mobbning i skolan. Rapport i samband med antimobbningsgruppens arbete. Institutionen för pedagogik vid Uppsala universitet

Sjölander, B. (1973) Mobbning. Psykologiska institutionen vid Uppsala universitet

Wramner, L. & Åhlander, E. (1973) Mobbning: Förebyggande åtgärder inom klassens ram, Pedagogiska institutionen vid Göteborgs universitet

Archives Radiohusets kortkatalog SAOB arkiv Statens ljud- och bildarkiv